nuevos avances en metodologías analíticas basadas en...

NUEVOS AVANCES EN METODOLOGÍAS ANALÍTICAS BASADAS EN TÉCNICAS

MINIATURIZADAS DE EXTRACCIÓN EN FASE SÓLIDA Y EN FASE LÍQUIDA

Elena Fernández Martínez

www.ua.es

www.eltallerdigital.com

Departamento de Química Analítica, Nutrición y Bromatología

e Instituto Universitario de Materiales

Facultad de Ciencias

NUEVOS AVANCES EN METODOLOGÍAS ANALÍTICAS BASADAS EN TÉCNICAS

MINIATURIZADAS DE EXTRACCIÓN EN FASE SÓLIDA Y EN FASE LÍQUIDA

Elena Fernández Martínez

Tesis presentada para la obtención del título de DOCTORA por la Universidad de Alicante con MENCIÓN DE DOCTORA

INTERNACIONAL

DOCTORADO EN CIENCIA DE MATERIALES

Dirigida por:

D. Antonio CanalsHernández

Catedrático de Universidad

Dña. Lorena Vidal Martínez

Profesora contratada doctora

Esta memoria es el resultado de años de trabajo y dedicación. Con estas palabras me gustaría dar las gracias a todos aquellos que, de una manera u otra, la han hecho posible:

A mis directores, Antonio Canals y Lorena Vidal, por ofrecerme la oportunidad de trabajar con ellos y vivir esta experiencia tan enriquecedora tanto a nivel personal como científico, por su orientación y ayuda durante todos estos años.

A mis compañeros/as del grupo de “Espectroscopía Atómica-Masas y Química Analítica en condiciones extremas”, por todos los buenos momentos compartidos, y porque los malos momentos fueron menos malos gracias a ellos.

A todos aquellos con los que he compartido laboratorio en algún momento de la tesis y de los que guardo entrañables recuerdos.

A mis padres, por su infinita generosidad y amor incondicional, por ser el pilar fundamental de lo que hoy en día soy. A mi hermana y a mi prima, porque sus visitas llegaron en el momento preciso y ayudaron a hacer más llevadera la distancia.

A Sáez, por su paciencia y comprensión, y porque su apoyo ha sido fundamental para llegar hasta aquí.

Por último, me gustaría agradecer a la Genaralitat Valenciana (VALi+d ACIF/2013/130) y al Ministerio de Educación (FPU13/03125) las ayudas predoctorales concedidas.

ÍNDICES

ÍNDICE GENERAL

I. Resumen/Abstract ..................................................................... 1

I.1. Resumen .............................................................................................. 3

I.2. Abstract ............................................................................................... 6

II. Introducción general ............................................................... 11

II.1. Preparación de la muestra .............................................................. 13

II.1.1. Técnicas miniaturizadas de extracción en fase sólida ................ 17

II.1.1.1. Introducción ............................................................................ 17

II.1.1.2. Extracción en fase sólida magnética ....................................... 23

II.1.2. Técnicas miniaturizadas de extracción en fase líquida ............... 31

II.1.2.1. Introducción ............................................................................ 31

II.1.2.1.1. Líquidos iónicos como fase extractante.......................... 43

II.1.2.2. Microextracción en gota en espacio de cabeza ....................... 49

II.1.2.3. Extracción con electromembrana ........................................... 54

II.1.2.4. Microextracción líquido-líquido dispersiva ........................... 63

II.1.2.4.1. Microextracción líquido-líquido dispersiva con formación

in-situ del líquido iónico .................................................................... 69

II.1.2.4.2. Microextracción líquido-líquido dispersiva en fase

reversa……………………………………………………………… 70

II.1.2.4.3. Microextracción líquido-líquido dispersiva asistida por

energía de ultrasonidos o por agitación con vórtex ........................... 71

II.1.3. Diseño estadístico de experimentos ............................................ 74

II.1.3.1. Diseño de Plackett-Burman .................................................... 77

II.1.3.2. Diseño central compuesto ....................................................... 79

II.2. Sistemas miniaturizados de detección............................................. 82

II.2.1. Electrodos serigrafiados .............................................................. 83

II.2.1.1. Técnicas electroanalíticas ........................................................ 87

II.2.1.1.1. Voltamperometría cíclica ................................................ 87

II.2.1.1.2. Técnicas voltamperométricas de impulsos ...................... 90

II.2.1.1.2.1. Voltamperometría diferencial de impulsos ................. 90

II.2.1.1.2.2. Voltamperometría de onda cuadrada ........................... 91

II.2.1.1.3. Voltamperometría de redisolución anódica ..................... 92

REFERENCIAS........................................................................................... 95

III. Parte experimental, resultados y discusión ........................ 117

Capítulo 1 ................................................................................................... 119

Zeolite/iron oxide composite as sorbent for magnetic solid-phase

extraction of benzene, toluene, ethylbenzene and xylenes from water

samples prior to gas chromatography-mass spectrometry ....................... 121

Capítulo 2 ................................................................................................... 149

Hydrophilic magnetic ionic liquid for headspace single-drop

microextraction of chlorobenzenes prior to thermal desorption gas

chromatography-mass spectrometry ........................................................ 151

Capítulo 3 ................................................................................................... 175

Complexation-mediated electromembrane extraction of highly polar basic

drugs - a fundamental study with catecholamines in urine as model

system…………………………………………………………………... 177

Capítulo 4 ................................................................................................... 203

III.4.1. Screen-printed electrode-based electrochemical detector coupled

with in-situ ionic-liquid-assisted dispersive liquid-liquid microextraction

for determination of 2,4,6-trinitrotoluene ............................................... 205

III.4.2. Screen-printed electrode based electrochemical detector coupled

with ionic liquid dispersive liquid-liquid microextraction and microvolume

back-extraction for determination of mercury in water samples. ……...227

III.4.3. Mercury determination in urine samples by gold-nanostructured

screen-printed carbon electrodes after vortex-assisted ionic liquid

dispersive liquid-liquid microextraction ................................................. 251

III.4.4. Rapid determination of hydrophilic phenols in olive oil by vortex-

assisted reversed-phase dispersive liquid-liquid microextraction and

screen-printed carbon electrodes ............................................................. 275

IV. Discusión general de los resultados/General discussion of the results ..................................................................................... 305

IV.1. Elementos comunes de la memoria .............................................. 308

IV.2. Elementos no comunes de la memoria ......................................... 316

IV.3. Common elements of the overview ............................................... 325

IV.4. Non-common elements of the overview ....................................... 333

V. Conclusiones generales/General conclusions ..................... 341

V.1. Conclusiones generales .................................................................. 343

V.2. General conclusions ....................................................................... 345

VI. Trabajos futuros/Future works ........................................... 349

VI.1. Trabajos futuros ............................................................................ 351

VI.2. Future works .................................................................................. 352

VII. ANEXOS .............................................................................. 355

VII.1. Comunicaciones en congresos........................................................ 357

VII.2. Otras publicaciones ........................................................................ 360

ÍNDICE DE FIGURAS

II. Introducción general

Figura II.1. Esquema general de las diferentes etapas incluidas en el proceso analítico total. ................................................................................................... 13

Figura II.2. Etapas principales de la extracción en fase sólida. ...................... 15

Figura II.3. Extracción líquido-líquido convencional. .................................... 15

Figura II.4. Distribución anual de las publicaciones aparecidas en el periodo 2000-2016 relacionadas con técnicas miniaturizadas de: (a) extracción en fase sólida; y (b) extracción en fase líquida. (Fuente: SciFinder®, enero 2017). ... 16

Figura II.5. Dispositivo comercial para la microextracción en fase sólida en fibra (Supelco ®). ............................................................................................. 18

Figura II.6. Modos de microextracción en fase sólida en tubo: (a) modo activo; y (b) modo pasivo. (Fuente: referencia [9]).......................................... 19

Figura II.7. Dispositivo comercial para la microextracción por sorción en barra agitadora (GERSTEL Twister ®). ........................................................... 20

Figura II.8. Dispositivo para la microextracción en película delgada. (Fuente: referencia [13]). ................................................................................................ 21

Figura II.9. Cronología de la aparición de las principales técnicas miniaturizadas de extracción en fase sólida. .................................................... 23

Figura II.10. Esquema general de la extracción en fase sólida magnética...... 24

Figura II.11. Nanopartículas magnéticas funcionalizadas utilizadas en la extracción en fase sólida magnética. (Fuente: referencia [22]). ....................... 25

Figura II.12. Diferentes configuraciones para la obtención de nanotubos de carbono magnéticos: (a) modificación endohédrica; y (b) modificación exohédrica. (Fuente: referencia [24]). .............................................................. 27

Figura II.13. Representación esquemática de una red metal-orgánica. (Fuente: referencia [31]). ................................................................................................ 28

Figura II.14. Estructura bidimensional y tridimensional de una zeolita. (Fuente: referencia [34]). ................................................................................... 29

Figura II.15. Esquema del primer sistema de microextracción basado en una gota. (Fuente: referencia [55]). .......................................................................... 31

Figura II.16. Sistemas de microextracción en gota diseñados por: (a) Liu y Daspugta; y (b) Jeannot y Cantwell. (Fuente: referencias [57,58]). ................. 32

Figura II.17. Sistema convencional para la microextracción en gota con inmersión directa. .............................................................................................. 33

Figura II.18. Microextracción líquido-líquido-líquido. ................................... 34

Figura II.19. Diferentes configuraciones de la microextracción en fibra hueca: (a) fibra en forma de “U”; y (b) fibra en forma de barra. ................................. 35

Figura II.20. Vista de la sección transversal de la fibra hueca sumergida en la muestra: (a) modalidad de dos fases; (b) modalidad de tres fases. ................... 36

Figura II.21. Sistema de microextracción en flujo continuo. (Fuente: referencia [66]). ................................................................................................. 37

Figura II.22. Microextracción en barra con disolvente. ................................... 38

Figura II.23. Microextracción en gota directamente suspendida. .................... 40

Figura II.24. Cronología de la aparición de las principales técnicas de microextracción en fase líquida. ........................................................................ 42

Figura II.25. Líquidos iónicos utilizados en técnicas de microextracción en fase líquida. Cationes: (a) metilimidazolio; (b) pirrolidinio; (c) piridinio; (d) tetraalquilamonio; (e) tetraalquilfosfonio. Aniones: (a) hexafluorofosfato; (b) tetrafluroroborato; (c) haluros; (d) bis[(trifluorometil)sulfonil]imida; (e) trifluoroacetato; (f) tris(pentafluoroetil)trifluorofosfato; (g) trifluorometanosulfonato. .................................................................................. 43

Figura II.26. Líquidos iónicos magnéticos utilizados en Química Analítica. Cationes: (a) metilimidazolio; (b) pirrolidinio; (c) piridinio; (d) tetraalquilamonio; (e) tetraalquilfosfonio. ......................................................... 46

Figura II.27. Esquema general de la microextracción líquido-líquido dispersiva basada en líquidos iónicos magnéticos. ........................................... 47

Figura II.28. Microextracción en gota en espacio de cabeza. ......................... 49

Figura II.29. Esquema del procedimiento de microextracción en espacio de cabeza en modo dinámico: (1) fase extractante en el interior de la jeringa; (2) introducción de la fase gaseosa en la jeringa; (3) expulsión de la fase gaseosa.50

Figura II.30. Esquema del sistema de transporte de analito entre las tres fases implicadas en la microextracción en gota en espacio de cabeza y sus áreas interfaciales (ga= gaseosa/acuosa; og= orgánica/gaseosa). (Fuente: referencia [98]). ................................................................................................................. 50

Figura II.31. Extracción con electromembrana. .............................................. 55

Figura II.32. Variación de la concentración del analito en la muestra, en la membrana líquida soportada y en la fase aceptora en función del tiempo según el modelo transitorio. (Fuente: referencia [103]).............................................. 59

Figura II.33. Extracción con electromembrana de especies ácidas y básicas de forma simultánea. ............................................................................................. 62

Figura II.34. Sistema miniaturizado de extracción con electromembrana. (Fuente: referencia [112]). ................................................................................ 63

Figura II.35. Esquema del procedimiento convencional de la microextracción líquido-líquido dispersiva. ................................................................................ 64

Figura II.36. Modificaciones más significativas de la microextracción líquido-líquido dispersiva. ............................................................................................. 66

Figura II.37. Ejemplos de tubos de extracción especiales diseñados para la microextracción líquido-líquido dispersiva con disolventes de baja densidad. (Fuente: referencia [117]). ................................................................................ 67

Figura II.38. Reacción de metátesis durante la microextracción líquido-líquido dispersiva con formación in-situ del líquido iónico. ........................................ 69

Figura II.39. Fenómeno de cavitación producido por las ondas de ultrasonidos………………………………………………………………….... 72

Figura II.40. Esquema general de un sistema o proceso en estudio mediante diseño estadístico de experimentos. (Fuente: referencia [145]). ...................... 74

Figura II.41. Electrodo serigrafiado comercial de la empresa Dropsens. ....... 84

Figura II.42. Representación esquemática del proceso de serigrafiado de electrodos. (Fuente: referencia [168]). .............................................................. 85

Figura II.43. (a) Señal de excitación triangular utilizada en voltamperometría cíclica. (b) Voltamperograma típico de una especie redox en disolución. ....... 88

Figura II.44. (a) Perfil concentración-distancia durante el barrido directo de potencial. (b) Perfil de concentración-distancia al invertir la dirección de barrido. (Fuente: referencias [174,175]). .......................................................... 89

Figura II.45. (a) Señal de excitación y (b) voltamperograma típico de la voltamperometría diferencial de impulsos. ....................................................... 91

Figura II.46. Señal de excitación en voltamperometría de onda cuadrada: Esw, amplitud de la onda cuadrada; y ΔEs, altura de escalón. ................................... 92

Figura II.47. Esquema del programa de potencial aplicado en la voltamperometría de redisolución anódica. ...................................................... 93

III. Parte experimental, resultados y discusión

Capítulo 1

Zeolite/iron oxide composite as sorbent for magnetic solid-phase extraction of

benzene, toluene, ethylbenzene and xylenes from water samples prior to gas

chromatography-mass spectrometry

Figure 1. Total ion chromatograms obtained in the SIM mode after MSPE of a blank (green line) and a standard solution spiked at 100 µg L-1 with target analytes (black line). ………………………………………………….…..…127

Figure S1. Photograph showing the magnetic attraction between ZSM-5/iron oxide composite and neodymium magnet. …………………...……………..128

Figure 2. High resolution XPS spectrum in the Fe 2p region for ZSM-5/iron oxide composite. …………………………………………………………….130

Figure S2. Nitrogen adsorption-desorption isotherms for pure ZSM-5 zeolite and ZSM-5/iron oxide composite. ……………………………………….….131

Figure S3. Pareto charts of the Plackett-Burman design obtained for: (a) benzene; (b) toluene; and (c) ethylbenzene. ...……………………………....134

Figure 3. Response surfaces of CCCD obtained by plotting the amount of sorbent vs. extraction time for: (a) benzene; (b) toluene; and (c) ethylbenzene. ……………………………………………………………………………….137

Figure 4. Study of sorbent reutilization using the same ZSM-5/iron oxide composite in twelve consecutive extractions. …………………………........139

Capítulo 2

Hydrophilic magnetic ionic liquid for headspace single-drop microextraction

of chlorobenzenes prior to thermal desorption gas chromatography-mass

spectrometry

Figure 1. Scheme of the overall proposed method. ………………………...156

Figure 2. Total ion chromatograms obtained in the SIM mode after HS-SDME of a blank (green line) and a standard solution spiked at 0.7 µg L-1 with target analytes (black line). The concentration of IS in both chromatograms was 0.5 µg L-1. ……………………………………………………………………….157

Figure S1. Magnetization of [Emin]2[Co(NCS)4] as a function of applied magnetic field at 300 K. ………………………………………………..…...159

Figure S2. Thermogravimetric curve of [Emin]2[Co(NCS)4]. ……………..159

Figure 3. Pareto charts of the Plackett-Burman design. ……………………162

Capítulo 3

Complexation-mediated electromembrane extraction of highly polar basic

drugs - a fundamental study with catecholamines in urine as model system

Figure 1. PBA complexation of diol groups. ……………………………….186

Figure 2. Effect of complexing reagent. Extraction conditions: concentration of analytes, 1 mg L-1; sample pH, 5; SLM, DEHPi; applied voltage, 50 V; extraction time, 10 min; acceptor solution, 20 mM formic acid. Error bars represent the standard deviation of three replicated analysis. ………………187

Figure 3. Effect of (a) PBA concentration, and (b) TFPBA concentration. Extraction conditions: concentration of analytes, 1 mg L-1; sample pH, 5; SLM, DEHPi; applied voltage, 50 V; extraction time, 10 min; acceptor solution, 20 mM formic acid. Error bars represent the standard deviation of three replicated analysis. ……………………………………………………………………..189

Figure S1. Extraction performance in urine sample using DEHPi as SLM in the absence of complexing reagent, with 3 mM PBA dissolved in sample solution and with 25 mM TFPBA dissolved in the SLM, respectively. Extraction conditions: concentration of analytes, 1 mg L-1; applied voltage, 50 V; extraction time, 10 min; acceptor solution, 20 mM formic acid. Error bars represent the standard deviation of three replicated analysis. ……...……….190

Figure S2. Effect of sample pH on EME of catecholamines. Extraction conditions: concentration of analytes, 1 mg L-1; SLM, DEHPi with 25 mM TFPBA; applied voltage, 50 V; extraction time, 10 min; acceptor solution, 20 mM formic acid. Error bars represent the standard deviation of three replicated analysis. ………………….………………………………………………….191

Figure 4. Effect of (a) applied voltage, and (b) extraction time. Extraction conditions: concentration of analytes, 1 mg L-1; sample pH, 4; SLM, DEHPi with 25 mM TFPBA; applied voltage, 50 V (if not indicated); extraction time, 10 min (if not indicated); acceptor solution, 20 mM formic acid. Error bars represent the standard deviation of three replicated analysis. ………………193

Figure 5. EME performance in urine sample under optimized conditions: (a) HPLC-UV chromatograms showing sample clean-up, and (b) system current profile. …………............................................................................................195

Capítulo 4

III.4.1. Screen-printed electrode-based electrochemical detector coupled with

in-situ ionic-liquid-assisted dispersive liquid-liquid microextraction for

determination of 2,4,6-trinitrotoluene

Figure 1. In-situ IL-DLLME coupled with SPGE. ……..…………………..211

Figure S1. Pareto charts of the multivariate optimization. (a) Plackett-Burman design. (b) Central composite design. AA and BB are the quadratic effects of amount of [Hmim][Cl] and sample volume, respectively. AB is the interaction effect between the two variables. …………………………………………...215

Figure 2. Response surface of CCD. ………………………………………..217

Figure 3. DPV curves of a blank and four TNT standards prepared in commercial [Hmim][NTf2]. ……………………………………………........218

Figure 4. DPV curves of a blank and a TNT standard of 80 μg L-1 in ultrapure water after in-situ IL-DLLME under optimum conditions. ............................219

III.4.2. Screen-printed electrode based electrochemical detector coupled with

ionic liquid dispersive liquid-liquid microextraction and microvolume back-

extraction for determination of mercury in water samples

Figure 1. In-situ IL-DLLME and ultrasound-assisted microvolume back-extraction coupled with SPCnAuE. ... ……………………………………....236

Figure S1. Pareto chart of the Plackett-Burman design. .…………………...239

Figure 2. Surfaces response of CCD design obtained by plotting: (a) HCl volume vs. back-extraction time (amount of [Hmim][Cl]: 40 mg); (b) HCl volume vs. amount of [Hmim][Cl] (back-extraction time: 10 min); (c) amount of [Hmim][Cl] vs. back-extraction time (HCl volume: 40 µL). ………..…...243

Figure 3. Square-wave voltammograms, after baseline correction, of a blank and mercury aqueous standards of 0.5, 1, 3, 5, 7 and 10 µg L-1 after in-situ IL-DLLME and back-extraction under optimum conditions. ………………….245


screen-printed carbon electrodes after vortex-assisted ionic liquid dispersive

liquid-liquid microextraction

Figure 1. Effect of sample volume. Extraction conditions: Amount of chelating agent, 40 µL of 2 mg mL-1 solution; IL amount, 0.06 g; vortex-assisted IL-DLLME extraction time, 3 min; volume of 4 M HCl acceptor solution during back-extraction, 10 µL; back-extraction time, 9 min. ………………….…...259

Figure 2. Effect of IL amount. Extraction conditions: Sample volume, 10 mL; amount of chelating agent, 40 µL of 2 mg mL-1 solution; vortex-assisted IL-DLLME extraction time, 3 min; volume of 4 M HCl acceptor solution during back-extraction, 10 µL; back-extraction time, 9 min. ……………………....260

Figure 3. Effect of vortex-assisted IL-DLLME extraction time. Extraction conditions: Sample volume, 10 mL; amount of chelating agent, 40 µL of 2 mg mL-1 solution; IL amount, 0.06 g; volume of 4 M HCl acceptor solution during back-extraction, 10 µL; back-extraction time, 9 min. …………....................261

Figure 4. Effect of back-extraction time. Extraction conditions: Sample volume, 10 mL; amount of chelating agent, 40 µL of 2 mg mL-1 solution; IL amount, 0.06 g; vortex-assisted IL-DLLME extraction time, 3 min; volume of 4 M HCl acceptor solution during back-extraction, 10 µL. ………………...262

Figure 5. Influence of the urine storage temperature on Hg2+ determination. …………………………………………………………………………….…263

Figure S1. Calibration curves obtained with aqueous standards and urine samples at different spiking levels: 5 µg L-1 (urine sample 1); 10 µg L-1 (urine sample 2); and 15 µg L-1 (urine sample 3). ………………………………….264

Figure S2. Square-wave voltammograms used to determine LOD emprirically in different urine samples. LOD was, respectively: (a) 0.5 µg L-1; (b) 1 µg L-1; and (c) 1.5 µg L-1. ………………………………………………….………..265

Figure 6. Square-wave voltammograms obtained during the analysis of a spiked urine sample at 5 µg L-1 by the standard addition method. Legend shows calibration standards concentration. ………………………………………...268


assisted reversed-phase dispersive liquid-liquid microextraction and screen-

printed carbon electrodes

Figure 1. Vortex-assisted RP-DLLME coupled with SPCEs. .......................280

Figure 2. (a) Cyclic voltammograms of 10 mg L-1 individual standards of hydrophilic phenols in aqueous 0.1 M HCl; and (b) DPV voltammograms of a 10 mg L-1 mixed standard solution in aqueous 0.1 M HCl (in red) and 0.1 M HCl aqueous extract after RP-DLLME of a VOO sample (in blue). ……….284

Figure S1. DPV voltammograms of caffeic acid at different concentrations of tyrosol. The concentration of caffeic acid was maintained at 10 mg L-1 in all the experiments. ……………………………………………………………..285

Figure S2. DPV voltammograms of tyrosol at different concentrations of caffeic acid. The concentration of tyrosol was maintained at 10 mg L-1 in all the experiments. ………………………………………………………….….286

Figure S3. Pareto charts of the Plackett-Burman design obtained for: (a) caffeic acid; and (b) tyrosol. ………………………………………………...288

Figure 3. Graphical comparison of the results obtained with the proposed method and the Folin-Ciocalteu method. ……………………………….…..297

Figure 4. Graphical representation of LDFs obtained during LDA. Groups are shown with different symbols: ROO, squares; VOO, triangles; EVOO, circles. Crosses mark the centroid of each group. ………………….……………….299

IV. Discusión general de los resultados/General discussion of the results

Figura IV.1. Esquema de los elementos diferenciadores entre los distintos trabajos de investigación presentados en esta memoria. ……….…………...317

Figure IV.2. Chart of the non-common elements between the different works included in this overview. …..……………………………………………….334

ÍNDICE DE TABLAS


Tabla II.1. Matriz de experimentos del diseño de Plackett-Burman para estudiar k=11 factores en N=12 experimentos. ................................................ 79

Tabla II.2. Matriz de experimentos de un diseño central compuesto para el estudio de k=2 factores en N=13 experimentos. .............................................. 81

III. Parte experimental, resultados y discusión

Capítulo 1

Zeolite/iron oxide composite as sorbent for magnetic solid-phase extraction of

benzene, toluene, ethylbenzene and xylenes from water samples prior to gas

chromatography-mass spectrometry

Table 1. Experimental factors and levels of the Plackett-Burman design. …132

Table S1. Matrix of experiments of the Plackett-Burman design. ………….133

Table 2. Factors, low, central and high levels, and star points used in CCCD. ……………………………………………………………………………….136

Table S2. Matrix of experiments of CCCD. ………………….…………….136

Table 3. Main analytical parameters of the proposed method (MSPE-GC-MS). ………………………………………………………………………..….......141

Table 4. Relative recoveries and CV values (in parentheses) obtained for the target analytes in the three studied real water samples. ………….…...……..142

Capítulo 2

Hydrophilic magnetic ionic liquid for headspace single-drop microextraction

of chlorobenzenes prior to thermal desorption gas chromatography-mass

spectrometry

Table S1. Experimental factors and levels of the Plackettt-Burman design. .160

Table S2. Matrix of experiments of the Plackett-Burman design. …………161

Table 1. Main analytical parameters of the proposed method (MIL-based HS-SDME-TD-GC-MS). ………………………………………………………..166

Table 2. Comparison of the developed MIL-based HS-SDME-TD-GC-MS method with previously reported methods to determine chlorobenzenes in water samples. ………………………………………………………………168

Table 3. Relative recoveries and CV values (in parentheses) obtained for the target analytes in the three studied real water samples. ……………………..170

Capítulo 4

III.4.1. Screen-printed electrode-based electrochemical detector coupled with

in-situ ionic-liquid-assisted dispersive liquid-liquid microextraction for

determination of 2,4,6-trinitrotoluene

Table 1. Experimental variables and levels of the Plackett-Burman design. ……………………………………………………………………………….214

Table 2. Variables, low and high levels, central and star points used in CCD. …………………………………………………………………………...…..216

Table 3. Characteristics of some electrochemical methods developed to determine TNT and other nitroaromatic explosives. ………….…………….221

III.4.2. Screen-printed electrode based electrochemical detector coupled with

ionic liquid dispersive liquid-liquid microextraction and microvolume back-

extraction for determination of mercury in water samples

Table 1. Comparison of different methods using SPELs for the determination of mercury in water samples. ………………………………………………..232

Table 2. Experimental factors and levels of the Plackettt-Burman design. ……………………………………………………………….………………238

Table 3. Factors, low and high levels, central and star points used in CCD design. ……………………………………………………….………………241

Table S1. Coefficients and p-values obtained in CCD design. ….………….241

Table 4. Relative recoveries and CV values (in parentheses) for the analysis of mercury in real-world water samples. ………………………………….…...247


screen-printed carbon electrodes after vortex-assisted ionic liquid dispersive

liquid-liquid microextraction

Table 1. Determination of Hg2+ in spiked human urine samples. ……….….268

Table 2. Comparison of different electrochemical methods for Hg2+ determination in urine samples. ……………………………………….…….270


assisted reversed-phase dispersive liquid-liquid microextraction and screen-

printed carbon electrodes

Table S1. Experimental factors and levels of the Plackettt-Burman design. ……………………………………………………………………………….287

Table 1. Main analytical parameters of the proposed method obtained with the matrix-matching calibration using sunflower oil as analyte-free sample. …..291

Table 2. Concentrations added and found, relative recoveries and coefficients of variation (in parentheses) during recovery studies in different olive oil samples. ……………………………………………………………………..292

Table 3. Caffeic acid equivalents and tyrosol content found in fifteen olive oil samples of different quality analyzed by the proposed method. ….....……...293

Table 4. Electrochemical methods for hydrophilic phenols determination in olive oil samples. …………….……………………………………………...295

IV. Discusión general de los resulados/General discussion of the results

Tabla IV.1. Datos representativos de las diferentes estrategias de preparación de la muestra investigadas en esta memoria. ………………………….….…309

Tabla IV.2. Principales parámetros analíticos de los métodos desarrollados en esta memoria. …………...………………………………..………………….312

Tabla IV.3. Recuperaciones relativas obtenidas durante el análisis de muestras reales con los métodos desarrollados en esta memoria. …………………….315

Tabla IV.4. Sistemas de separación-detección y tiempos de análisis de los métodos desarrollados en los tres primeros capítulos de la memoria. ……...323

Tabla IV.5. Electrodos, técnicas electroanalíticas y tiempos de análisis de los métodos desarrollados en el Capítulo 4. ……………………...……..............323

Table IV.6. Representative parameters of the sample preparation strategies developed in the present overview. ................................................................326

Table IV.7. Main analytical figures of merit of the methods developed in the present overview. ………………………….………………….……..............329

Table IV.8. Relative recoveries obtained during the analysis of real samples with the methods developed in this overview. .……………..……................332

Table IV.9. Separation-detection systems and analysis time of the analytical methods included in the first three chapters of this overview. …………..….339

Table IV.10. Electrodes, electroanalytical techniques and analysis time of the analytical methods included in Chapter 4. ……..……………………..…….339

GLOSARIO DE TÉRMINOS Y ACRÓNIMOS

ANOVA Análisis de la varianza Analysis of variance

APDC Pirrolidín ditiocarbamato amónico Ammonium pyrrolidine dithiocarbamate

ASV Voltamperometría de redisolución anódica Anodic stripping voltammetry

BTEX Benceno, tolueno, etilbenceno y xileno Benzene, toluene, ethylbenzene and xylene

CCCD Diseño central compuesto circunscrito Circumscribed central composite design

CCD Diseño central compuesto Central composite design

CE Contraelectrodo Counter electrode

CFME Microextracción en flujo continuo Continuous-flow microextraction

CNTs Nanotubos de carbono Carbon nanotubes

CV Coeficiente de variación Coefficient of variation

DA Dopamina Dopamine

DBB Dibromobenceno Dibromobenzene

DCB Diclorobenceno Dichlorobenzene

DEHP Fosfato de di(2-etilhexilo) Di(2-ethylhexyl) phosphate

DEHPi Fosfito de bis(2-etilhexilo) Bis(2-ethylhexyl) phosphite

DI-SDME Microextracción en gota con inmersión directa Direct immersion single-drop microextraction

DLLME Microextracción líquido-líquido dispersiva Dispersive liquid-liquid microextraction

DPV Voltamperometría diferencial de impulsos Differential pulse voltammetry

DSDME Microextracción en gota directamente suspendida Directly suspended droplet microextraction

DSPME Microextracción en fase sólida dispersada Dispersive solid-phase microextraction

E Epinefrina Epinephrine

EF Factor de enriquecimiento Enrichment factor

ENB 1-etil-2-nitrobenceno 1-ethyl-2-nitrobenzene

[Emin]2[Co(NCS)4] Tetraisotiocianatocobaltato (II) de 1-etil-3-metilimidazolio

1. 1-ethyl-3-methylimidazolium tetraisothiocyanatocobaltate (II)

EME Extracción con electromembrana Electromembrane extraction

EPA Agencia de Protección Medioambiental Environmental Protection Agency

EVOO Aceite de oliva virgen extra Extra virgin olive oil

FID Detector de ionización en llama Flame ionization detector

GC Cromatografía de gases Gas chromatography

HF-LPME Microextracción en fibra hueca Hollow fiber liquid-phase microextraction

[Hmim][Cl] Cloruro de 1-hexil-3-metilimidazolio 1-hexyl-3-methylimidazolium chloride

[Hmim][NTf2] Bis[(trifluorometil)sulfonil]imida de 1-hexil-3-metilimidazolio 1-hexyl-3-methylimidazolium bis[(trifluoromethyl)sulfonyl]imide

HPLC Cromatografía de líquidos de alto rendimiento High performance liquid chromatography

HS-SDME Microextracción en gota en espacio de cabeza Headspace single-drop microextraction

IL Líquido iónico Ionic liquid

IMS Espectrometría de movilidad iónica Ion mobility spectrometry

In-situ IL-DLLME

Microextracción líquido-líquido dispersiva con formación in-situ del líquido iónico In-situ ionic liquid formation dispersive liquid-liquid

microextraction

IPNB 1-(isopropil)-4-nitrobenceno 1-isopropyl-4-nitrobenzene

LC Cromatografía de líquidos Liquid chromatography

LDA Análisis discriminante lineal Linear discriminant analysis

LDF Función discriminante lineal Linear discriminant function

LIBS Espectrometría de plasmas inducidos por láser Laser-induced breakdown spectrometry

LiNTf2 Bis[(trifluorometil)sulfonil]imida de litio Lithium bis[(trifluoromethyl)sulfonyl]imide

LLE Extracción líquido-líquido Liquid-liquid extraction

LLLME Microextracción líquido-líquido-líquido Liquid-liquid-liquid microextraction

LOD Límite de detección Limit of detection

LOQ Límite de cuantificación Limit of quantification

LPME Microextracción en fase líquida Liquid-phase microextraction

MEPS Microextracción en jeringa empaquetada Microextraction in a packed syringe

MIL Líquido iónico magnético Magnetic ionic liquid

MNPs Nanopartículas magnéticas Magnetic nanoparticles

MOF Red metal-orgánica Metal-organic framework

MS Espectrometría de masas Mass spectrometry

MSPE Extracción en fase sólida magnética Magnetic solid-phase extracion

NE Norepinefrina Norepinephrine

NPOE 2-nitrofenil octil éter 2-nitrophenyl octylether

PBA Ácido fenilborónico Phenylboronic acid

PDMS Polidimetilsiloxano Polydimethylsiloxane

PeCB Pentaclorobenceno Pentachlorobenzene

r Coeficiente de correlación Correlation coefficient

RE Electrodo de referencia Reference electrode

ROO Aceite de oliva refinado Refined olive oil

RP-DLLME Microextracción líquido-líquido dispersiva en fase reversa Reversed-phase dispersive liquid-liquid microextraction

SBME Microextracción en barra con disolvente Solvent bar microextraction

SBSE Extracción por sorción en barra agitadora Stir bar sorptive extraction

SD Desviación estándar Standard deviation

SFOD Solidificación de la gota orgánica flotante Solidification of the floating organic drop

SFODME Microextracción con solidificación de la gota orgánica flotante Solidification of the floating organic drop microextraction

SIM Monitoreo selectivo de iones Selected ion monitoring

SLM Membrana líquida soportada Supported liquid membrane

SPCE Electrodo serigrafiado de carbono Screen-printed carbon electrode

SPCnAuE Electrodo serigrafiado de carbono modificado con nanopartículas de oro Gold nanostructured screen-printed carbon electrode

SPE Extracción en fase sólida Solid-phase extraction

SPEL Electrodo serigrafiado Screen-printed electrode

SPGE Electrodo serigrafiado de grafito Screen-printed graphite electrode

SPME Microextracción en fase sólida Solid-phase microextraction

SWASV Voltamperometría de onda cuadrada de redisolución anódica Square-wave anodic stripping voltammetry

SWV Voltamperometría de onda cuadrada Square-wave voltammetry

TCB Triclorobenceno Trichlorobenzene

TC-IL-DLLME

Microextracción líquido-líquido dispersiva basada en un líquido iónico y con control de temperatura

Temperature-controlled ionic liquid dispersive liquid-liquid

microextraction

TeCB Tetraclorobenceno Tetrachlorobenzene

TEHP Fosfato de tris(2-etilhexilo) Tri(2-ethylhexyl) phosphate

TFME Microextracción en película delgada Thin-film microextraction

TFPBA Ácido 4-(trifluorometil)fenilborónico 4-(trifluoromethyl)phenylboronic acid

TG Termogravimétrico Thermogravimetric

TNT 2,4,6-trinitrotolueno 2,4,6-trinitrotoluene

TD Desorción térmica Thermal desorption

UA-DLLME

Microextracción líquido-líquido dispersiva asistida por ultrasonidos Ultrasound-assisted dispersive liquid-liquid microextraction

UHPLC Cromatografía de líquidos de ultra alto rendimiento Ultra-high performance liquid chromatography

UPD Deposición a subpotencial Underpotential deposition

VA-DLLME Microextracción líquido-líquido dispersiva asistida por vórtex Vortex-assisted dispersive liquid-liquid microextraction

VOO Aceite de oliva virgen Virgin olive oil

WE Electrodo de trabajo Working electrode

WHO Organización mundial de la salud World Health Organization

XPS Espectroscopía fotoelectrónica de rayos X X-ray photoelectron spectroscopy

µ-SPE Microextracción en fase sólida con membrana porosa protectora Porous membrane-protected micro-solid-phase extraction

I. Resumen/Abstract

Resumen

3

I.1. Resumen La demanda de métodos analíticos rápidos, sencillos, económicos,

respetuosos con el medio ambiente y que permitan realizar análisis descentralizados y en tiempo real es cada vez mayor. Para cumplir con estos objetivos, la miniaturización de las metodologías y dispositivos utilizados resulta fundamental. Cuando se habla de la miniaturización de un método de análisis conviene considerar por separado dos partes del proceso analítico total: (a) la preparación de la muestra; y (b) la separación-detección (i.e., proceso de medida). La miniaturización de la preparación de la muestra se ha convertido en una tendencia dominante dentro de la Química Analítica actual, dando lugar a nuevas técnicas de microextracción en fase sólida y en fase líquida. Estas técnicas miniaturizadas reducen el consumo de muestra y reactivos, los tiempos de análisis, costes y riesgos, con todas las ventajas que estos avances suponen. Por otro lado, la miniaturización de los sistemas de separación-detección no ha alcanzado el mismo grado de desarrollo que el de la preparación de la muestra, debido a la complejidad de la instrumentación analítica y de la tecnología necesaria para tal fin. Por tanto, hace falta realizar un mayor esfuerzo investigador para miniaturizar las técnicas instrumentales existentes o utilizar las actualmente disponibles en formato miniaturizado pero que han sido escasamente utilizadas tras una etapa de microextracción. En este sentido, los sensores electroquímicos son una opción especialmente interesante y prometedora por tratarse de una tecnología en formato miniaturizado muy desarrollada y de bajo coste.

La presente Tesis Doctoral posee dos objetivos fundamentales. El primero de ellos consiste en desarrollar y evaluar nuevas estrategias para la miniaturización de la preparación de la muestra, así como avanzar en la investigación de las ya existentes. El segundo objetivo está centrado en la utilización y evaluación de los electrodos serigrafiados como sistemas de detección alternativos para las técnicas de microextracción, desarrollando así métodos analíticos en los que se utilizan sistemas miniaturizados tanto en la preparación de la muestra como en la etapa de detección.

Resumen

4

Para cumplir con los objetivos propuestos, la presente memoria recoge siete trabajos de investigación distribuidos en cuatro capítulos, los cuales se describen brevemente a continuación:

Capítulo 1: en este primer capítulo se investiga la extracción en fase sólida magnética de benceno, tolueno, etilbenceno y xileno desde muestras de agua, utilizando un nuevo sólido sorbente basado en una zeolita sintética modificada con nanopartículas magnéticas.

Capítulo 2: en segundo lugar se evalua la utilización de un líquido iónico magnético para la extracción en gota en espacio de cabeza de clorobencenos contenidos en muestras de agua, previamente a su separación-detección mediante desorción térmica-cromatografía de gases-espectrometría de masas.

Capítulo 3: en tercer lugar se lleva a cabo un estudio fundamental sobre la extracción con electromembrana de analitos básicos de elevada polaridad como son las catecolaminas (dopamina, epinefrina y norepinefrina).

Capítulo 4: este último capítulo agrupa cuatro trabajos de investigación debido a las características comunes que poseen. En todos ellos se investiga una técnica de microextracción líquido-líquido dispersiva y su combinación con electrodos serigrafiados como sistemas de detección electroquímica. En cada trabajo se abordan problemas concretos y se proponen nuevas metodologías analíticas para la determinación de distintos analitos en muestras de diversa naturaleza. - Apartado III.4.1: se presenta por primera vez la combinación de

una técnica de microextracción líquido-líquido dispersiva con electrodos serigrafiados en la etapa de medida. Se utiliza un líquido iónico como fase extractante cuyas propiedades electroquímicas permiten la determinación directa del analito modelo, 2,4,6-trinitrotolueno, tras la etapa de extracción.

- Apartado III.4.2: el objetivo de este segundo trabajo es el desarrollo de un método analítico para la determinación de mercurio en muestras de agua. El método propuesto utiliza, en la etapa de preparación de la muestra, la técnica de microextracción

Resumen

5

líquido-líquido dispersiva con formación in-situ del líquido iónico seguida de una re-extracción asistida por ultrasonidos. En la etapa de detección se utilizan electrodos serigrafiados de carbono modificados con nanopartículas de oro.

- Apartado III.4.3: este trabajo se ha centrado en la determinación de mercurio en muestras de orina y está basado en parte en el método desarrollado en el apartado anterior. No obstante, se introducen ciertas modificaciones en la preparación de la muestra ya que en este caso, tanto la microextracción líquido-líquido dispersiva como la re-extracción son asistidas por vórtex.

- Apartado III.4.4: este último trabajo recoge la determinación de

fenoles hidrofílicos en aceites de oliva. Se propone un método basado en una microextracción líquido-líquido dispersiva en fase reversa asistida por vórtex seguida de la detección electroquímica con electrodos serigrafiados de carbono.

La presente Tesis Doctoral ha sido realizada en el grupo de investigación de Espectroscopía Atómica-Masas y Química Analítica en condiciones extremas del Departamento de Química Analítica, Nutrición y Bromatología, y del Insituto Universitario de Materiales de la Universidad de Alicante. Además, parte de la investigación ha sido desarrollada en el grupo de NanoBioAnálisis del Departamento de Química Física y Analítica de la Universidad de Oviedo, bajo la supervisión del catedrático D. Agustín Costa-García, y en el grupo de Bioanálisis del Departamento de Química Farmacéutica de la Universidad de Oslo, bajo la supervisión del doctor D. Stig Pedersen-Bjergaard. Debido a que la presente Tesis Doctoral opta a la Mención Internacional, algunos apartados de esta memoria han sido redactados en inglés, con el propósito de cumplir con la normativa vigente.

Abstract

6

I.2. Abstract The demand for fast, simple, economical and environmentally friendly

analytical methods allowing decentralized and real-time analysis is continuously increasing. The use of miniaturized methodologies and devices is essential in order to fulfill all these objectives. When discussing the miniaturization of an analytical method, two parts of the total analytical process should be considered separately: (a) sample preparation; and (b) separation-detection (i.e., measurement process). The miniaturization of sample preparation has become a dominant trend in the current Analytical Chemistry, leading to new solid-phase and liquid-phase microextraction techniques. These miniaturized techniques reduce the consumption of sample and reagents, analysis time, costs and risks, thus providing remarkable advantages. On the contrary, separation and detection systems have not achieved miniaturization to the same extent as sample preparation, due to complex analytical instrumentation and the technology needed for this purpose. Therefore, greater efforts are required to miniaturize traditional instrumental techniques or to use those currently available in miniaturized format, but being scarcely used after microextraction techniques. In this regard, electrochemical sensors are a very interesting and promising option having an inexpensive and well-developed miniaturized technology.

This Doctoral Thesis possesses two main objectives. The first one concerns the development and evaluation of new strategies for the miniaturization of sample preparation, as well as the improvement of other previously developed. The second objective is focused on the use and evaluation of screen-printed electrodes as alternative detection systems for microextraction techniques, thus developing analytical methods in which miniaturized systems are used both in sample preparation and in detection step.

In order to fulfill the proposed objectives, seven research works are included in this overview, distributed into four different chapters, which are briefly described below:

• Chapter 1: magnetic solid-phase extraction of benzene, toluene, ethylbenzene and xylene from water samples is investigated, using a

Abstract

7

new solid sorbent based on a synthetic zeolite decorated with magnetic nanoparticles.

• Chapter 2: a new magnetic ionic liquid is evaluated for the headspace single-drop microextraction of chlorobenzenes from water samples prior to thermal desorption gas chromatography-mass spectrometry.

• Chapter 3: a fundamental study is carried out on the electromembrane extraction of highly polar basic compounds such as catecholamines (dopamine, epinephrine and norepinephrine).

• Chapter 4: four research works are included in this last chapter due to their common characteristics. In all of them, a dispersive liquid-liquid microextraction technique coupled to screen-printed electrodes as electrochemical detection systems are investigated. Following this new concept, new analytical methods are proposed for the determination of different analytes in samples of different nature. - Section III.4.1: dispersive liquid-liquid microextraction technique

is coupled to screen-printed electrodes for the first time. An ionic liquid is used as extractant phase whose electrochemical properties allow the direct determination of target model analyte (i.e., 2,4,6-trinitrotoluene) after microextraction step.

- Section III.4.2: a novel approach is presented, whereby gold nanostructured screen-printed carbon electrodes are combined with in-situ ionic liquid formation dispersive liquid-liquid microextraction and ultrasound-assisted microvolume back-extraction for the determination of mercury in water samples.

- Section III.4.3: a method is proposed to determine mercury in urine samples partly based on the developed method in the previous section. However, some modifications are introduced during sample preparation since vortex-assisted ionic liquid dispersive liquid-liquid microextraction and vortex-assisted microvolume back-extraction are employed in this case.

- Section III.4.4: this last work reports the determination of hydrophilic phenols in olive oil samples. The proposed method is based on vortex-assisted reversed-phase dispersive liquid-liquid

Abstract

8

microextraction followed by the electrochemical detection with screen-printed carbon electrodes.

This Doctoral Thesis has been carried out in the research group of Atomic-Mass Spectroscopy and Analytical Chemistry under extreme conditions of the Department of Analytical Chemistry, Nutrition and Food Sciences, and the University Institute of Materials of the University of Alicante. Part of the research was developed in the group of NanoBioAnalysis of the Department of Physical and Analytical Chemistry of the University of Oviedo, under the supervision of Full Professor Mr. Agustín Costa-García, and in the group of Bioanalysis of the Department of Pharmaceutical Chemistry of the University of Oslo, under the supervision of Dr. Mr. Stig Pedersen-Bjergaard. Because this Doctoral Thesis aims for International Mention, some sections of this overview have been written in English, with the purpose of fulfilling the current regulations.

Introducción general

13

II.1. Preparación de la muestra El objetivo principal de todo proceso analítico es la obtención de

información (bio)química de calidad sobre materiales o sistemas de diversa naturaleza (química, bioquímica o biológica) cuya interpretación permita dar respuesta a un determinado problema científico-técnico o socio-económico [1]. La resolución de dichos problemas mediante la obtención de información (bio)química requiere de una metodología sistemática y concreta que en Química Analítica recibe el nombre de proceso analítico total. Este proceso consta de una serie de etapas que, de forma general, pueden esquematizarse tal y como se muestra en la Figura II.1 [2,3].

Figura II.1. Esquema general de las diferentes etapas incluidas en el proceso analítico total.

Identificación del problema científico-técnico o socio-económico

Identificación de la información (bio)analítica necesaria para su resolución

Planificación de la estrategia y procedimiento analítico a seguir

Toma de muestra

Preparación de la muestra

Medida

Tratamiento y evaluación de los datos obtenidos

Interpretación de los resultados y conclusiones

¿satisfactorio?

SÍ

Informe final de los resultados

NO

Medidas correctivas


14

El esquema general mostrado en la Figura II.1 no es en absoluto inalterable. Por una lado, la frontera entre las diferentes etapas no siempre es clara y, por otro, algunas de ellas pueden ser obviadas en función del problema abordado o de la información solicitada [3].

De las diferentes etapas que constituyen el proceso analítico total, la preparación de la muestra juega un papel fundamental. En esta etapa se consigue dar a la muestra el estado físico adecuado para la técnica de análisis seleccionada, eliminar posibles interferencias, concentrar o diluir el analito, etc. Es evidente que el mejor método de análisis estaría basado en la medida directa de la muestra sin que ésta necesitara ser tratada. Sin embargo, las técnicas instrumentales utilizadas, la existencia de matrices complejas y la necesidad de alcanzar límites de detección cada vez más bajos hacen que en la inmensa mayoría de las ocasiones la etapa de preparación de la muestra siga siendo necesaria e imprescindible.

Entre la gran cantidad de técnicas existentes para la preparación de la muestra, las técnicas de extracción son las más utilizadas. Estas técnicas están basadas en el reparto de los analitos entre la muestra y una fase extractante, pudiendo ser ambas de naturaleza sólida, líquida o gaseosa [4]. No obstante, la presente memoria está centrada en el tratamiento de muestras líquidas mediante la extracción de los analitos objeto de estudio a una fase extractante sólida o líquida.

En la extracción en fase sólida (solid-phase extraction, SPE) convencional se utilizan sorbentes dispuestos en cartuchos (tubos) o discos (membranas) a través de los cuales se hace circular la muestra líquida. El procedimiento general de esta técnica de extracción consta de cuatro etapas principales (Figura II.2) [4]: (i) acondicionamiento del cartucho o disco; (ii) paso de la muestra a través del cartucho o disco; (iii) elución de las especies interferentes mediante un disolvente adecuado; y (iv) elución de los analitos de interés a un extracto limpio y concentrado.


15

Figura II.2. Etapas principales de la extracción en fase sólida.

Por otro lado, la extracción líquido-líquido (liquid-liquid extraction, LLE) convencional consiste en la mezcla íntima de la muestra con un disolvente adecuado (i.e., inmiscible y capaz de extraer las especies de interés pero no las impurezas que las acompañan), de manera que los analitos se distribuyen entre ambas fases hasta alcanzar el equilibrio. Generalmente se utiliza un embudo de decantación que se agita durante un tiempo adecuado y que, tras la extracción, se deja reposar produciéndose la separación espontánea de las fases inmiscibles por diferencia de densidades (Figura II.3).

Figura II.3. Extracción líquido-líquido convencional.

Agitación manual del embudo de decantación

durante la extracción

Reposo y separación espontánea de las fases

Acondicionamiento

Analito Especies interferentes

Adición de la muestra

Lavado de las especies interferentes

Elución de los analitos


16

Tanto la SPE como la LLE presentan ciertos inconvenientes, principalmente relacionados con un consumo elevado de reactivos y disolventes, en muchos casos caros y/o tóxicos, y tiempos largos de extracción. Por tanto, estas técnicas tradicionales resultan caras, tediosas y poco respetuosas con el medio ambiente. Como consecuencia, una de las tendencias más significativas en los últimos años dentro de la Química Analítica ha consistido en la miniaturización de las técnicas de extracción. Prueba de esta tendencia es el creciente número de publicaciones anuales relacionadas con técnicas miniaturizadas de extracción en fase sólida y en fase líquida que se muestra en la Figura II.4.

Figura II.4. Distribución anual de las publicaciones aparecidas en el periodo 2000-2016 relacionadas con técnicas miniaturizadas de: (a) extracción en fase sólida; y (b) extracción en fase líquida. (Fuente: SciFinder®, enero 2017).

0200400600800

100012001400

Nº p

ublic

acio

nes

Año

0

100

200

300

400

500

Nº p

ublic

acio

nes

Año

(b)

(a)


17

El aumento del interés de los investigadores por las técnicas miniaturizadas de extracción es debido a las importantes ventajas que éstas han aportado dentro del campo de la preparación de la muestra. Tanto las técnicas miniaturizadas de extracción en fase sólida como las de extracción en fase líquida reducen el consumo de muestra y reactivos, la generación de residuos, los tiempos de análisis y los costes. De este modo, se ha conseguido desarrollar métodos analíticos más sostenibles y menos perjudiciales para el medio ambiente y la salud del operador.

II.1.1. Técnicas miniaturizadas de extracción en fase sólida

II.1.1.1. Introducción

La miniaturización de la SPE tuvo como punto de partida el año 1990, cuando Pawliszyn y Arthur presentaron por primera vez la microextracción en fase sólida (solid-phase microextraction, SPME) [5]. Esta técnica nació con el fin de eliminar, o al menos reducir, dos de las principales limitaciones de la SPE: el consumo relativamente elevado de disolventes y los tiempos largos de extracción [5].

La SPME cuenta con un dispositivo basado en una fibra de sílice fundida, químicamente inerte, estable a altas temperaturas y recubierta de un polímero sorbente (Figura II.5). El proceso de extracción consta de dos etapas principales: (i) reparto de los analitos entre la muestra y el polímero sorbente; y (ii) desorción de los analitos retenidos en la fibra mediante temperatura o con un disolvente orgánico. De este modo, la SPME en fibra es un procedimiento sencillo, con posibilidades de automatización y de acoplamiento on-line con dispositivos de medida basados en la cromatografía de gases (gas

chromatography, GC). Además, otras de las principales ventajas que presenta esta técnica es que puede utilizarse con todos los tipos de muestra (i.e., sólida, líquida o gaseosa) y para la extracción de analitos con propiedades y polaridades muy variadas, gracias al gran número de recubrimientos comercialmente disponibles en la actualidad [6].

Técnicas miniaturizadas de extracción en fase sólida

18

Figura II.5. Dispositivo comercial para la microextracción en fase sólida en fibra (Supelco ®).

A pesar de las importantes ventajas que supuso la aparición de la SPME en

fibra, esta técnica todavía cuenta con ciertas limitaciones. Entre los inconvenientes más importantes podemos destacar la cantidad limitada de fase sorbente, que afecta directamente a la cantidad de analito extraído, la fragilidad de la fibra, los posibles efectos de memoria a la hora de reutilizarla, su elevado coste y la dificultad del acoplamiento on-line con otras técnicas instrumentales como la cromatografía de líquidos (liquid chromatography, LC). Estos inconvenientes han provocado la aparición con el tiempo de nuevas técnicas miniaturizadas de SPE destinadas a superarlos.

En el año 1997, Eisert y Pawliszyn presentaron la SPME en tubo [7]. Esta técnica cuenta con dos configuraciones principales [8]. La primera de ellas utiliza un tubo capilar abierto por sus dos extremos por el que se hace circular la muestra para la retención de los analitos en el material sorbente que cubre las paredes internas de dicho capilar (Figura II.6a). La segunda configuración consiste en un tubo capilar cerrado por uno de sus extremos de manera que la muestra es succionada y expulsada repetidamente de su interior (Figura II.6b). En ambos casos, tras la extracción los analitos son eluídos normalmente

Émbolo

Aguja

Fibra recubierta con el polímero sorbente


19

utilizando un disolvente. Aunque la SPME en tubo fue originalmente propuesta para su acoplamiento on-line con la LC [7], posteriormente ha sido utilizada junto a otras técnicas de separación como la GC o la electroforesis capilar [8]. Además, se han desarrollado dispositivos de extracción alternativos en los que el material sorbente rellena el tubo capilar en lugar de recubrir sus paredes internas [8].

Figura II.6. Modos de microextracción en fase sólida en tubo: (a) modo activo; y (b) modo pasivo. (Fuente: referencia [9]).

La principal limitación que presenta la SPME en tubo es el bloqueo del

capilar debido a muestras con matrices sucias, por lo que éstas deben ser filtradas antes de someterse al proceso de extracción.

El año 1999 fue el año de presentación de dos nuevas técnicas en las que se proporcionaban propiedades magnéticas a la fase extractante: la extracción por sorción en barra agitadora (stir bar sorptive extraction, SBSE) [10] y la extracción en fase sólida magnética (magnetic solid-phase extracion, MSPE) [11]. La primera de ellas está basada en una barra agitadora magnética recubierta de vidrio y por un polímero sorbente de polidimetilsiloxano (polydimethylsiloxane, PDMS) (Figura II.7). Este dispositivo se introduce en la muestra y se hace girar con una placa agitadora. Después de la extracción, la

Muestra Muestra

Tubo capilar

Sorbente

(a) (b)


20

barra magnética es retirada manualmente y los analitos son desorbidos térmicamente o con un pequeño volumen de disolvente. Aunque la SBSE posee una mayor capacidad de extracción que la SPME en fibra debido a la mayor cantidad de material sorbente disponible, esta técnica no ha alcanzado una aceptación tan amplia debido al limitado número de recubrimientos comercialmente disponibles.

Figura II.7. Dispositivo comercial para la microextracción por sorción en barra agitadora (GERSTEL Twister ®).

Por otro lado, la MSPE utiliza como fase extractante un sorbente sólido

con propiedades magnéticas o al que se le han conferido dichas propiedades. Este sólido es dispersado directamente en la muestra y tras la extracción, es separado de la misma mediante la aplicación de un campo magnético externo. Seguidamente, los analitos retenidos en el sólido magnético son eluidos utilizando un disolvente apropiado.

Como se ha comentado anteriormente, una de las principales limitaciones de la SPME en fibra es la pequeña cantidad de sorbente disponible para la extracción de los analitos. Con el fin de superar este inconveniente, Bruheim et al. propusieron en el año 2003 la técnica de microextracción en película delgada (thin-film microextraction, TFME) [12]. La TFME utiliza como fase extractante una lámina delgada, generalmente de PDMS, con una elevada relación área/volumen (Figura II.8). Con esta configuración, la cantidad disponible de fase extractante es mayor que en la SPME en fibra, lo que da lugar a una mayor capacidad de extracción. Además, al existir una la elevada

PDMS Barra magnética

Recubrimiento de vidrio


21

superficie de contacto entre las fases, los tiempos de extracción no se ven aumentados.

Figura II.8. Dispositivo para la microextracción en película delgada. (Fuente: referencia [13]).

En el año 2004, Abdel-Rehim propuso la técnica de microextracción en jeringa empaquetada (microextraction in a packed syringe, MEPS) [14]. Esta técnica consiste en una miniaturización de la SPE convencional en el sentido más estricto de la definición. La MEPS utiliza como dispositivo de extracción una jeringa de entre 100 y 250 µL rellena de unos pocos mg de material sorbente empaquetado. Al igual que en la SPE convencional, la muestra se hace circular a través del sólido sorbente donde los analitos quedan retenidos para, posteriormente, ser eluídos con un disolvente apropiado. La MEPS cuenta con dispositivos comerciales basados en los mismos tipos de sorbentes que los utilizados en los cartuchos de SPE tradicionales. Sin embargo, la técnica miniaturizada cuenta con la ventaja de que los tiempos de extracción y el volumen de muestra necesarios para alcanzar un alto factor de enriquecimiento se ven enormemente reducidos [6].

Barra magnética agitadora

Película delgada de polímero sorbente

Muestra

Soporte de Teflón


22

La extracción directa de analitos desde matrices complejas con sólidos en suspensión ocasiona ciertos inconvenientes en técnicas como la SPME en fibra (e.g., daño o rotura de la fibra) o la SPME en tubo (e.g., bloqueo del capilar). Como solución a este problema, en el año 2006 Basheer et al. propusieron la microextracción en fase sólida con membrana porosa protectora (porous

membrane-protected micro-solid-phase extraction, µ-SPE) [15]. Este dispositivo de extracción consiste en una pequeña bolsa porosa hecha de polipropileno en la que se introduce y protege el material sorbente. Esta bolsa es sumergida directamente en la muestra de modo que la membrana porosa impide la entrada del material particulado y evita la contaminación de la fase extractante por los componentes de la matriz.

En el año 2009, Tsai et al. acuñaron por primera vez el término de microextracción en fase sólida dispersada (dispersive solid-phase

microextraction, DSPME) [16]. Los autores presentaron esta nueva técnica como una miniaturización de la extracción en fase sólida dispersada basada en los conocidos QuEChERS [17,18]. En la DSPME, unos pocos µg o mg de sorbente sólido son introducidos directamente en la muestra y dispersados en la misma mediante agitación manual o utilizando dispositivos adicionales como el vórtex o la energía de ultrasonidos [19]. Después de la extracción, las fases son separadas por centrifugación o filtración y los analitos son desorbidos térmicamente o con un disolvente para su determinación. La dispersión de la fase extractante en la muestra da lugar a una mayor área superficial activa del sorbente y a un contacto íntimo entre las fases por lo que la extracción es más eficiente y el tiempo para alcanzar el equilibrio disminuye.

A modo de resumen, la Figura II.9 muestra la cronología de la aparición de las principales técnicas miniaturizadas de SPE introducidas en este apartado.


23

Figura II.9. Cronología de la aparición de las principales técnicas miniaturizadas de extracción en fase sólida.

De todas las técnicas descritas anteriormente, la MSPE será desarrollada con mayor detalle a continuación, ya que esta técnica ha sido objeto de estudio en el Capítulo 1 de la presente memoria.

II.1.1.2. Extracción en fase sólida magnética

La MSPE está basada en la dispersión en la muestra de un sorbente sólido con propiedades magnéticas. Dicha dispersión puede lograrse mediante agitación manual o con dispositivos adicionales como el vórtex [20] o la energía de ultrasonidos [21]. Tras la extracción, el sorbente sólido enriquecido es separado de la muestra con la ayuda de un imán, de manera que ésta puede ser descartada fácilmente por decantación o utilizando una pipeta Pasteur. Una vez se ha eliminado la muestra, el sorbente magnético es dispersado en un disolvente apropiado para la elución de los analitos. Finalmente, se utiliza de nuevo el imán para atraer y separar al sorbente sólido del eluyente o eluato enriquecido, y este último es recogido para su análisis. La Figura II.10

muestra un esquema general del procedimiento.

1990

1997

1998

1999

2003

2004

2005

2006 SPME en tubo

- SBSE - MSPE

TFME

µ-SPE

SPME en fibra

MEPS

2009 DSPME


24

Figura II.10. Esquema general de la extracción en fase sólida magnética.

La MSPE permite una rápida separación de las fases mediante la

aplicación de un campo magnético externo, evitando así etapas lentas de centrifugación o filtración y minimizando la manipulación de la muestra. Otras de las principales ventajas que presenta esta técnica es una elevada capacidad de extracción con tiempos cortos de operación, debido a la dispersión de la fase extractante y a su elevada área superficial disponible. Por otro lado, la MSPE puede ser utilizada en muestras con matrices sucias o partículas en suspensión ya que se eliminan los problemas asociados a otras técnicas miniaturizadas de extracción en fase sólida, como la rotura de la fibra (SPME en fibra), el bloqueo del tubo capilar (SPME en tubo) o de la jeringa empaquetada (MEPS). Además, las impurezas o sólidos en suspensión de una muestra son normalmente diamagnéticos de manera que no interactúan con el sorbente magnético durante la separación de las fases [22–24].

Los sorbentes sólidos magnéticos más utilizados en la MSPE son las nanopartículas magnéticas (magnetic nanoparticles, MNPs) funcionalizadas [22,23]. Las MNPs (e.g., Fe2O3, Fe3O4, MnFe2O4, CoFe2O4) poseen una elevada relación área-volumen y un comportamiento superparamagnético de manera que pueden ser atraídas por un imán pero no poseen magnetismo residual una vez se elimina el campo magnético externo. Uno de los principales problemas asociados a las partículas de tamaño nanométrico es su inestabilidad

Dispersión del sorbente sólido

magnético

Elución

Adición del disolvente

Eliminación de la fase acuosa

Imán Separación de

las fases

Imán Imán Retirada del

disolvente para su análisis


25

y la consiguiente tendencia a formar aglomerados. Además, muchas nanopartículas metálicas son fácilmente oxidables en contacto con el aire, lo que conlleva la pérdida de sus propiedades magnéticas. Por tanto, para la MSPE con MNPs funcionalizadas es esencial en primer lugar proteger y estabilizar las nanopartículas mediante algún tipo de recubrimiento (Figura II.11). Este recubrimiento sirve a la vez de base para la modificación superficial con diferentes compuestos o materiales que interaccionan con los analitos y permiten su extracción (Figura II.11). Dicha modificación superficial le otorga una elevada versatilidad a la técnica y ofrece la posibilidad de obtener sistemas altamente selectivos para un determinado analito de interés [22,23].

Figura II.11. Nanopartículas magnéticas funcionalizadas utilizadas en la extracción en fase sólida magnética. (Fuente: referencia [22]).

Nanotubos de carbono

Celulosa

Alquil C18

Surfactantes

Quitosano

Nanopartículas metálicas

Polímero de impresión molecular

Anticuerpos

Aptámeros

Líquidos iónicos Otros

Núcleo magnético

Superficie modificada Recubrimiento

Dióxido de silicio

Óxido de aluminio

Dióxido de titanio

Zirconio

Sin recubrimiento

Otros


26

Por otro lado, la MSPE también puede utilizar diferentes tipos de materiales sorbentes que, por naturaleza propia, no poseen propiedades magnéticas pero que pueden ser modificados con el fin de otorgarles dichas propiedades. Entre estos materiales podemos destacar los de base carbonosa (e.g., grafeno, óxido de grafeno, nanotubos de carbono), las redes metal-orgánicas (metal-organic frameworks, MOFs) y las zeolitas.

El grafeno posee una estructura plana en dos dimensiones formada por átomos de carbono con hibridación sp2 y una elevada área superficial. Su carga π deslocalizada le permite formar fuertes interacciones π-π con anillos aromáticos, lo que le convierte en un excelente sorbente para analitos que incluyen un anillo de benceno en su estructura. Por otro lado, en la ruta de síntesis del grafeno a partir del grafito, es posible obtener una especie intermedia denominada óxido de grafeno. Este material es rico en grupos radicales epoxi, carboxilo e hidroxilo que son capaces de formar puentes de hidrógeno o interacciones electrostáticas con los grupos polares de especies orgánicas y con iones metálicos. Tanto el grafeno como el óxido de grafeno han sido utilizados como fase extractante en diferentes técnicas de SPME [19,25–27]. Para la MSPE, la magnetización del grafeno puede conseguirse mediante la coprecipitación sobre su superficie de sales de Fe(II) y Fe(III) en medio básico [28]. Por otro lado, los nanocompuestos basados en óxido de grafeno/Fe3O4 pueden ser fácilmente obtenidos mediante la simple interacción electrostática de las MNPs (carga superficial positiva) con el material carbonoso, el cual posee una carga superficial negativa consecuencia de la ionización de grupos carboxílicos y fenólicos [23,29]. La modificación del grafeno o del óxido de grafeno con MNPs permite obtener sorbentes en los que se combinan sinérgicamente las excelentes propiedades extractantes del material carbonoso con las ventajas que proporciona el magnetismo para su manipulación [30].

Los nanotubos de carbono (carbon nanotubes, CNTs) son estructuras consistentes en una única lámina de grafeno enrollada en forma de tubo (nanotubos monocapa), o bien en varios de estos tubos con diámetros crecientes, dispuestos alrededor de un eje común (nanotubos multicapa). Los CNTs poseen una excelente capacidad de extracción debido a su elevada área


27

superficial. Además, la superficie de estos materiales puede ser fácilmente modificada introduciendo grupos funcionales que interaccionen específicamente con el analito de interés. La magnetización de los CNTs para su utilización como sorbentes en MSPE se lleva a cabo mediante la incorporación, bien endohédrica o bien exohédrica (Figura II.12), de diferentes metales u óxidos metálicos, principalmente MNPs de Fe3O4 [24]. Esta incorporación puede realizarse en dos pasos o en un único paso. El procedimiento de dos pasos incluye la síntesis por separado de la MNPs y su posterior mezcla y auto-ensamblaje con los CNTs. Alternativamente, los CNTs magnéticos pueden obtenerse mediante la precipitación directa e in-situ de las MNPs sobre la superficie del material carbonoso.

Figura II.12. Diferentes configuraciones para la obtención de nanotubos de carbono magnéticos: (a) modificación endohédrica; y (b) modificación exohédrica. (Fuente: referencia [24]).

La magnetización de los CNTs ha permitido superar el principal

inconveniente de la utilización de este material como fase sólida extractante, que reside en la dificultad de separarlo y recuperarlo una vez dispersado en una disolución acuosa. Los CNTs magnéticos han sido satisfactoriamente utilizados para la extracción tanto de analitos orgánicos como de iones metálicos desde muestras complejas de muy diversa naturaleza (e.g., biológicas, de alimentos, etc.) [24].

(a) (b)


28

Los MOFs son compuestos de coordinación formados por la unión de centros metálicos o clústers a través de ligandos orgánicos para generar estructuras que se extienden en el espacio en varias dimensiones (Figura II.13).

Figura II.13. Representación esquemática de una red metal-orgánica. (Fuente: referencia [31]).

Esta nueva clase de materiales porosos ha generado un gran interés durante los últimos años debido a las múltiples aplicaciones que pueden encontrar, como la adsorción de gases o la catálisis heterogénea. Además, su estructura porosa uniforme y de escala nanométrica, su elevada área superficial y su estabilidad térmica han propiciado un vasto número de aplicaciones como sorbentes en técnicas de SPME [32]. Para la MSPE, los MOFs magnéticos pueden obtenerse a través de diferentes procedimientos [31]. El más sencillo de ellos consiste en el autoensamblaje de las MNPs sobre la superficie exterior del MOF mediante la dispersión de una mezcla acuosa de ambos materiales [33].

Por último, otro de los materiales utilizados como fase extractante en MSPE han sido las zeolitas. Las zeolitas son aluminosilicatos hidratados cristalinos formados por una red tridimensional de tetraedros de (SiO4)4- y (AlO4)5- interconectados por sus vértices mediante átomos de oxígeno (Figura II.14). La presencia de cada tetraedro de (AlO4)5- genera una carga negativa global que es compensada por cationes, normalmente alcalinos o alcalinotérreos (e.g., Na+, K+, Ca2+), situados en los canales o poros que forman

Ligando orgánico

Iones metálicos o clústers

Red metal-orgánica (MOF)


29

dicha red. Estos cationes pueden ser sustituidos con facilidad por otros, otorgando a las zeolitas una gran capacidad de intercambio iónico. Además, estos materiales poseen una elevada área superficial y capacidad de adsorción, pudiendo actuar como tamices moleculares al ser capaces de discriminar entre moléculas de diferente tamaño y forma. Finalmente, otra propiedad importante inherente a las zeolitas es su actividad catalítica debido a los grupos hidroxilo terminales que actúan como ácidos de Brönsted [34].

Figura II.14. Estructura bidimensional y tridimensional de una zeolita. (Fuente: referencia [34]).

Las zeolitas se encuentran disponibles en la naturaleza en forma de minerales, pero también pueden ser sintetizadas en el laboratorio o a nivel industrial. Actualmente, la Asociación Internacional de Zeolitas reconoce la existencia de casi 200 estructuras diferentes, incluyendo tanto las naturales como las sintéticas [35]. La diferente morfología (i.e., forma y tamaño de poros y canales) y composición química de las zeolitas les confieren propiedades muy diversas, pudiendo elegir la más adecuada para una aplicación específica. En general, las estructuras con una elevada relación Si/Al poseen mayor estabilidad térmica, mayor hidrofobicidad y acidez, pero menor capacidad de intercambio iónico [36]. Las zeolitas han sido utilizadas principalmente como

M+ M

+

poros

cavidad central


30

catalizadores en la industria petroquímica y para la limpieza de aguas contaminadas, debido a su capacidad de adsorción de compuestos orgánicos y de intercambio iónico de metales pesados [34]. Además, el creciente interés por la síntesis de nuevas zeolitas ha promovido aplicaciones innovadoras en células solares, microelectrónica, medicina y sensores holográficos [37]. A pesar del uso habitual y establecido de las zeolitas en los campos mencionados, su aplicación en el análisis químico ha sido mucho más limitada hasta la fecha [34]. No obstante, la enorme versatilidad de estos materiales y su bajo coste han captado recientemente el interés de los analistas. Como consecuencia, las zeolitas han sido propuestas para modificar la superficie electródica en sensores y biosensores electroquímicos, como fase estacionaria en LC y como fase sólida en diferentes técnicas de extracción (e.g., SPE, SPME, DMSPE, TFME) [34,37–45]. En cuanto a la MSPE, recientemente se han presentado dos trabajos de investigación en los que se utilizan zeolitas magnéticas como fase extractante [46,47]. En el primero de ellos, la zeolita natural clinoptilolita es modificada mediante la deposición superficial de MNPs de Fe3O4 obtenidas previamente por un método electroquímico. Esta zeolita es utilizada para la extracción de ftalatos desde muestras de agua. En el segundo trabajo, la síntesis y la magnetización de la zeolita ZSM-5 se llevan a cabo de manera simultánea a partir de Fe3O4, SiO2 y NaAlO2. Posteriormente, la zeolita magnética es modificada con calix[n]arenos que permiten la extracción de antioxidanes fenólicos mediante interacciones π-π, hidrofóbicas y puentes de hidrógeno. Los autores de este trabajo realizaron un estudio comparativo sobre la capacidad de extracción de la zeolita magnética sin modificar y modificada, demostrando que la segunda exhibía un mejor rendimiento.

La MSPE es considerada una técnica sencilla y rápida que ha aportado numerosas ventajas dentro del campo de la preparación de la muestra. Las investigaciones dirigidas a aumentar el conocimiento sobre este tipo de microextracción y nuevas fases extractantes magnéticas son muy numerosas en la actualidad, como pone de manifiesto la gran cantidad de revisiones bibliográficas aparecidas en los últimos años [22–24,30,31,48–53].


31

II.1.2. Técnicas miniaturizadas de extracción en fase líquida

II.1.2.1. Introducción

Las técnicas de microextracción en fase líquida (liquid-phase

microextraction, LPME) son consideradas técnicas miniaturizadas de LLE en las que los analitos son extraídos y concentrados en una fase extractante con un volumen igual o inferior a 100 µL [54]. Aunque la mayoría de los autores establecen el año 1996 como el año de nacimiento de la LPME, fue en 1995 cuando Liu y Dasgupta presentaron el primer método de análisis basado en una extracción en gota [55]. En dicho método, los analitos (i.e., NH3 y SO2) eran extraídos, desde una muestra gaseosa, en una gota de unos 5 µL de una disolución acuosa suspendida al final de un tubo capilar (Figura II.15). En ese mismo año, Cardoso y Dasgupta sustituyeron la fase extractante en forma de gota por una película delgada (14-57 µL) para la extracción de NO2 [56].

Figura II.15. Esquema del primer sistema de microextracción basado en una gota. (Fuente: referencia [55]).

Muestra (gas)

Bomba de aspiración

Soporte para el capilar

Entrada/salida fase extractante

Tubo capilar

Gota

Técnicas miniaturizadas de extracción en fase líquida

32

En el año 1996, Liu y Dasgupta presentaron un nuevo sistema de extracción en el que 1.3 µL de cloroformo eran suspendidos en el interior de una gota acuosa de mayor tamaño que contenía los analitos y que era continuamente renovada mediante un sistema de bombeo (sistema drop-in-

drop) (Figura II.16a) [57]. Casi simultáneamente, Jeannot y Cantwell utilizaron 8 µL de n-octano suspendidos al final de una varilla de Teflón y sumergidos en una muestra acuosa agitada por una barra magnética (Figura II.16b) [58].

Figura II.16. Sistemas de microextracción en gota diseñados por: (a) Liu y Daspugta; y (b) Jeannot y Cantwell. (Fuente: referencias [57,58]).

Un año después de su propuesta inicial (Figura II.16b), Jeannot y Cantwell presentaron una configuración alternativa en la que la gota de disolvente extractante era suspendida en la punta de la aguja de una jeringa [59]. En ese momento nació lo que hoy en día se conoce como la microextracción en gota con inmersión directa (direct immersion single-drop

microextraction, DI-SDME) (Figura II.17).

Fase acuosa

Cloroformo

Tubo de PTFE

Gota orgánica

Gota acuosa

Varilla de Teflón

Vial

Barra agitadora

Tapón

Fase orgánica

Fase acuosa

(a) (b)


33

Figura II.17. Sistema convencional para la microextracción en gota con inmersión directa.

Hasta el año 1998, los diferentes sistemas de LPME presentados habían considerado la participación de dos fases: la muestra acuosa y el disolvente orgánico extractante. Fue entonces cuando Ma y Cantwell propusieron por primera vez un sistema de extracción de tres fases conocido como microextracción líquido-líquido-líquido (liquid-liquid-liquid microextraction, LLLME) [60]. En dicho sistema, los analitos son extraídos desde la muestra acuosa a un disolvente orgánico y posteriormente son re-extraídos a una fase acuosa aceptora final. El disolvente orgánico, menos denso que el agua, actúa por tanto como interfase entre las dos disoluciones acuosas. El primer sistema de LLLME empleó un anillo de Teflón para confinar una fina película de fase orgánica en la que se suspendía una gota de fase aceptora acuosa con la ayuda de una jeringa [60]. Sin embargo, modificaciones posteriores de la técnica eliminaron la utilización del anillo de Teflón y la fase orgánica es colocada directamente en la muestra (Figura II.18) [61,62].

Jeringa

Muestra (5-30 mL)


Gota (1-3 µL)

Placa agitadora


34

Figura II.18. Microextracción líquido-líquido-líquido. La LLME es apropiada para la extracción de compuestos orgánicos

ionizables con propiedades ácido-base. Esto hace que el pH de ambas disoluciones acuosas sea un fator crítico en esta técnica de microextracción. Por ejemplo, para la extracción de especies básicas, el pH de la muestra debe ser ajustado por encima del pK de los analitos con el fin de asegurar la supresión de su ionización, disminuir su solubilidad en la fase acuosa y facilitar así su transferencia al disolvente orgánico. Al mismo tiempo, la fase aceptora es acidificada (pH˂pK) de manera que los analitos son ionizados en la interfase disolvente orgánico/fase acuosa aceptora, aumentando su solubilidad en ésta última. Siguiendo un razonamiento análogo, la extracción de especies ácidas requiere el ajuste del pH de la muestra por debajo del pK de los analitos y la utilización de una fase aceptora basificada (pH˃pK).

La suspensión de la fase extractante en la punta de la aguja de una jeringa presenta ciertos problemas de inestabilidad, riesgo de desprendimiento y limitaciones a la hora de agitar la muestra. Este hecho no pasó desapercibido por los investigadores y en 1999, Pedersen-Bjergaard y Rasmussen propusieron

Placa agitadora


Muestra

Fase orgánica Gota de fase

aceptora acuosa

Jeringa


35

la técnica de microextracción en fibra hueca (hollow fiber liquid-phase

microextraction, HF-LPME) [63]. La HF-LPME está basada en la extracción de los analitos desde la muestra a una fase aceptora (e.g., disolución acuosa o disolvente orgánico) protegida en el interior de una fibra hueca polimérica en cuyos poros se encuentra inmovilizado un disolvente orgánico formando una membrana líquida soportada (supported liquid membrane, SLM). La Figura II.19 muestra las dos configuraciones principales para llevar a cabo la HF-LPME según la disposición y morfología de la fibra. En la primera de ellas (Figura II.19a), la fibra en forma de “U” es sumergida en la muestra y sus extremos son conectados a sendas jeringas, una para la inyección de la fase aceptora y otra para su retirada. Esta configuración fue presentada en el primer trabajo de HF-LPME [63]. No obstante, la configuración mostrada en la Figura II.19b se ha convertido en la más utilizada actualmente. En ella, la fibra en forma de barra es sellada por uno de sus extremos mientras que el otro es conectado a una jeringa utilizada tanto para la inyección de la fase aceptora como para su retirada una vez finalizada la extracción. La longitud de la fibra en forma de barra es menor que la de la fibra en forma de “U” por lo que el volumen de la fase aceptora es más reducido y, por tanto, los factores de enriquecimiento que se alcanzan son mayores.

Figura II.19. Diferentes configuraciones de la microextracción en fibra hueca: (a) fibra en forma de “U”; y (b) fibra en forma de barra.

(a)

Muestra

Fase aceptora Fibra hueca

(SLM)

Jeringa


Fibra hueca (SLM)

Inyección de la fase aceptora

Muestra

Retirada de la fase aceptora

Fase aceptora

Jeringa Jeringa


(b)


36

Por otro lado, según participen dos o tres fases en el proceso de extracción, existen dos modalidades diferentes de HF-LPME. En la HF-LPME de dos fases, el mismo disolvente orgánico es inmovilizado en los poros de la fibra hueca e introducido en el interior de la misma (Figura II.20a). El proceso de extracción incluye la transferencia de los analitos desde la muestra acuosa hasta la SLM y su posterior difusión hacia el seno de la fase aceptora orgánica. Por tanto, esta modalidad es adecuada para la extracción de analitos hidrofóbicos con elevados coeficientes de partición entre el disolvente orgánico y la muestra.

Figura II.20. Vista de la sección transversal de la fibra hueca sumergida en la muestra: (a) modalidad de dos fases; (b) modalidad de tres fases.

La HF-LPME de tres fases está basada en la transferencia de los analitos

desde la muestra hasta la SLM, su difusión a través de la misma y su posterior re-extracción a una fase acuosa aceptora final (Figura II.20b). El proceso de extracción está afectado tanto por el coeficiente de partición de los analitos entre la fase orgánica y la muestra como por el coeficiente de partición entre la fase aceptora y la fase orgánica. Por consiguiente, la HF-LPME de tres fases es adecuada para la extracción de analitos con propiedades ácido/base cuya solubilidad en las diferentes fases puede ser modulada mediante cambios en el pH, tal y como se ha descrito anteriormente para la LLLME. Por otro lado, la extracción de especies ionizables polares con baja afinidad por la SLM es poco

Fase aceptora orgánica en el interior de la fibra hueca

Disolvente orgánico inmovilizado en los poros de la fibra (SLM)

(a)

Muestra Muestra

Fase aceptora acuosa en el interior de la fibra hueca

Muestra

Disolvente orgánico inmovilizado en los poros de la fibra (SLM)

(b)

Muestra


37

eficiente mediante simple difusión pasiva. Estas especies requieren de un transporte activo para alcanzar la fase aceptora final. Para ello, se han utilizado especies transportadoras o carriers que interaccionan con los analitos polares formado pares iónicos hidrofóbicos que son fácilmente transferidos a la fase orgánica [64].

En el año 2000, Liu y Lee introdujeron una nueva técnica de LPME a la que dieron el nombre de microextracción en flujo continuo (continuous-flow

microextraction, CFME) [65]. La CFME está basada en la exposición de la fase extractante a una muestra acuosa en movimiento gracias a un sistema de bombeo. La extracción se lleva a cabo en una cámara de vidrio donde se expone la gota de disolvente suspendida al final de un tubo capilar [65]. El paso de la muestra en un flujo continuo favorece la extracción, alcanzando altos factores de enriquecimiento con tiempos cortos de extracción. Después de la extracción, la gota es recogida con una jeringa e inyectada en el sistema de análisis seleccionado. Alternativamente, la gota puede ser expuesta a la corriente de la muestra a través de una jeringa (Figura II.21) [66]. De esta manera, la misma jeringa es utilizada para sostener la fase extractante durante la extracción e inyectarla posteriormente en el sistema de detección.

Figura II.21. Sistema de microextracción en flujo continuo. (Fuente: referencia [66]).

Otra modificación de la CFME consiste en el bombeo de la muestra en un flujo continuo recirculante [67]. Es decir, a la salida de la cámara de extracción, la muestra es dirigida de nuevo al recipiente que la contiene y

Bomba

Flujo de la muestra

Tubo de Teflón

Fase orgánica

Cámara de extracción Jeringa


38

sometida repetidamente al proceso de extracción. Esta recirculación disminuye considerablemente el consumo de muestra, pudiéndose realizar extracciones desde 1 ó 2 mL de la misma.

Todas las técnicas de LPME descritas hasta el momento cuentan con la existencia de un contacto directo entre la muestra y la fase extractante, ya sea a través de una gota sumergida o de una SLM. El contacto directo entre ambas fases conlleva ciertos problemas de inestabilidad provocados por matrices sucias, la agitación de la muestra o la solubilidad del disolvente. Por esta razón, en el año 2001 dos grupos de investigación diferentes propusieron de manera casi simultánea la microextracción en gota en espacio de cabeza (headspace

single-drop microextraction, HS-SDME), en la que la gota de disolvente es expuesta en el espacio de cabeza situado por encima de la muestra evitando el contacto directo entre ambas fases [68,69].

Entre los años 2004 y 2006, la HF-LPME fue objeto de varias modificaciones. Por un lado, Jiang y Lee propusieron la microextracción en barra con disolvente (solvent bar microextraction, SBME) [70]. En la SBME, un trozo de fibra hueca rellena con la fase extractante es sellada por ambos extremos y sumergida directamente en la muestra sin utilizar ningún tipo de soporte (i.e., jeringa) durante la extracción (Figura II.22) [70]. De este modo, la fibra se mueve libremente por la muestra como consecuencia de la agitación magnética de la misma, pudiendo obtener mayores factores de enriquecimiento que con las configuraciones mostradas en la Figura II.19 [70]. Posteriormente, se extrae la fibra de la muestra, se corta un extremo de la misma y se retira la fase extractante con una jeringa para ser analizada.

Figura II.22. Microextracción en barra con disolvente.

Muestra

Fibra hueca (SLM)

Fase aceptora



39

Por otro lado, Pedersen-Bjergaard y Rasmussen utilizaron un campo

eléctrico para asistir la HF-LPME en la modalidad de tres fases, dando lugar a la técnica de microextracción conocida como extracción con electromembrana (electromembrane extraction, EME) [71].

Hasta el año 2006, las técnicas de LPME aparecidas utilizaban diferentes tipos de soportes (i.e., jeringa, fibra hueca, tubo capilar) para sostener la fase extractante durante la extracción. No obstante, en aquel momento se presentaron dos nuevas técnicas en las que se obviaba la utilización de cualquier soporte y la fase extractante era colocada directamente en la muestra: la microextracción líquido-líquido dispersiva (dispersive liquid-liquid

microextraction, DLLME) y la microextracción en gota directamente suspendida (directly suspended droplet microextraction, DSDME). La DLLME está basada en la dispersión de la fase extractante en el seno de la muestra y en una etapa posterior de centrifugación para separar las fases [72]. Por el contrario, en la DSDME la fase extractante conserva su integridad en forma de gota y se mantiene flotando sobre la superficie de la muestra acuosa sometida a agitación [73]. Una vez finalizada la extracción, la gota es recogida con una jeringa para ser analizada (Figura II.23). Algunos autores han señalado que el tamaño de la barra magnética agitadora puede afectar a la forma de la gota dificultando su retirada después de la extracción. Por este motivo, el procedimiento convencional de la DSDME ha sido modificado introduciendo un sistema de agitación por rotación del vial que contiene la muestra [74]. La ausencia de la barra magnética agitadora permite formar gotas más estables y reproducibles proporcionando una mayor precisión al método.


40

Figura II.23. Microextracción en gota directamente suspendida.

La retirada de la gota tras la DSDME es una etapa complicada que requiere cierto cuidado y habilidad por parte del operador. Como solución a este problema, en el año 2007 se presentó la microextracción con solidificación de la gota orgánica flotante (solidification of the floating organic drop

microextraction, SFODME) [75]. En esta nueva técnica, los pasos iniciales son idénticos a los de la DSDME. Es decir, la gota de disolvente extractante es directamente suspendida en la superficie de la muestra sin la utilización de ningún tipo de soporte. Sin embargo, la SFODME conlleva una etapa posterior de solidificación de la gota suspendida con el fin de facilitar su retirada [75]. Para tal fin, una vez finalizada la extracción, el vial es introducido en un baño de hielo durante unos 5 minutos aproximadamente. Seguidamente, la fase orgánica solidificada es transferida a otro vial vacío utilizando una pequeña espátula. Una vez allí, la gota de disolvente solidificado funde rápidamente a temperatura ambiente y puede ser retirada para su inyección en el sistema de análisis seleccionado. La SFODME requiere de disolventes orgánicos que

Jeringa

Muestra Barra magnética

agitadora

Gota

Placa agitadora

Extracción Retirada de la gota tras la extracción


41

posean un punto de fusión cercano a la temperatura ambiente (entre 10 y 30 ºC), además de una baja solubilidad en agua y baja volatilidad. A pesar de las ventajas en el modo de operar que presenta esta técnica, la separación de las fases no siempre es satisfactoria ya que resulta prácticamente inevitable la presencia de pequeñas cantidades de agua en la fase extractante, que son eliminadas tras una etapa adicional de centrifugado. Por tanto, la SFODME cuenta con un gran número de etapas que la convierten en una técnica tediosa y poco práctica.

A modo de resumen, la cronología de la aparición de las principales técnicas de LPME descritas en este apartado se muestra esquemáticamente en la Figura II.24. Todas estas técnicas han experimentado numerosas modificaciones desde sus respectivos nacimientos. Por ejemplo, se han propuesto fases extractantes alternativas a los disolventes orgánicos tradicionales (e.g., líquidos iónicos, surfactantes, etc.), se han utilizado distintos modos de energía o radiación para calentar o dispersar las fases (e.g., microondas, ultrasonidos, etc.), se han desarrollado novedosas combinaciones con sistemas de detección y algunos procedimientos han sido automatizados [54,76]. De forma general, los investigadores han tendido a acuñar nuevos términos y acrónimos para nombrar las diferentes modificaciones introducidas. Esta tendencia ha tenido como consecuencia la aparición de más de 100 acrónimos diferentes relacionados con las técnicas de LPME [54]. Este hecho ha sido criticado explícitamente en una publicación reciente ya que pone de manifiesto la falta de consenso dentro de la comunidad científica y dificulta enormemente la búsqueda bibliográfica del investigador [54].

De todas las técnicas mostradas en la Figura II.24, la HS-SDME, la EME y la DLLME serán desarrolladas con mayor profundidad en apartados individuales de esta memoria ya que son las técnicas de microextracción objeto de estudio en el trabajo de investigación que aquí se presenta.


42

Figura II.24. Cronología de la aparición de las principales técnicas de microextracción en fase líquida.

1995

1996

1997

1998 2000

1999 2001

2004

2005

2006

2007

- Sistema drop-in-drop - Gota suspendida en un tubo de Teflón

DI-SDME

LLLME

HF-LPME

CFME

HS-SDME

SBME - EME - DLLME - DSDME

SFODME Primer sistema de extracción en gota


43

II.1.2.1.1. Líquidos iónicos como fase extractante

Los disolventes orgánicos como el tolueno, el hexano o el cloroformo han sido habitualmente utilizados como fase extractante en técnicas de LPME. Sin embargo, estos disolventes tradicionales son volátiles y generalmente tóxicos por lo que la comunidad científica creyó en la necesidad de sustituirlos por otros disolventes alternativos con el fin de desarrollar metodologías menos perjudiciales para el medio ambiente y la salud del operador. Dentro de este contexto, en el año 2003 se presentó por primera vez la utilización de un líquido iónico (ionic liquid, IL) como fase extractante para la SDME [77]. Los ILs son sales líquidas a temperatura ambiente o a temperaturas cercanas a ésta, siendo 100 ºC su punto de fusión máximo [78,79]. En la mayoría de los casos, los ILs están formados por cationes orgánicos y aniones orgánicos o inorgánicos. El gran tamaño de uno o de los dos iones que forman la sal, la consecuente asimetría que esto provoca y la separación de cargas, impiden la organización de los ILs en estructuras cristalinas haciendo que se encuentren en estado líquido a bajas temperaturas. En la Figura II.25 se muestran algunos de los cationes y aniones más frecuentes en los ILs utilizados en técnicas de LPME.

Figura II.25. Líquidos iónicos utilizados en técnicas de microextracción en fase líquida. Cationes: (a) metilimidazolio; (b) pirrolidinio; (c) piridinio; (d) tetraalquilamonio; (e) tetraalquilfosfonio. Aniones: (a) hexafluorofosfato; (b) tetrafluroroborato; (c) haluros; (d) bis[(trifluorometil)sulfonil]imida; (e) trifluoroacetato; (f) tris(pentafluoroetil)trifluorofosfato; (g) trifluorometanosulfonato.

Líquidos iónicos

NN RH3C

N

R2R1

N R4

R1

R3

R2

N

R

P R4

R1

R3

R2

Cl- Br-BF4- PF6

-

C

O-F3C

O

S

O

O

CF3-O

NSS

O

O

O

O

CF3F3C

P

F

C2F5C2F5

F F

C2F5

BF4- PF6

-Cl- Br-

Cationes Aniones

(a) (b) (c)

(d) (e)

(a) (b) (c) (d)

(e) (f) (g)


44

Los ILs poseen propiedades únicas entre las que se encuentran una buena estabilidad térmica y química, baja volatilidad e inflamabilidad, viscosidad y densidad variables, conductividad iónica intrínseca, amplia ventana electroquímica, miscibilidad selectiva en agua o disolventes orgánicos y buena extractabilidad de especies orgánicas, metálicas y organometálicas [78,79]. Además, uno de los aspectos a destacar en los ILs es que variando el catión o el anión podemos modificar significativamente sus propiedades físicas y químicas. Como consecuencia, la combinación de la gran variedad de cationes y aniones existentes da lugar a un enorme número de ILs con propiedades diferentes. Esto proporciona un amplio intervalo de posibilidades entre las que se puede seleccionar el IL más apropiado para una aplicación concreta.

La utilización de los ILs como fase extractante en las técnicas de LPME ha aportado importantes ventajas [80]. Por ejemplo, en la SDME se han conseguido superar en parte los problemas relacionados con la inestabilidad de la gota suspendida en la punta de la jeringa. Al mismo tiempo, la inmiscibilidad de determinados ILs en agua permite utilizarlos en la DI-SDME empleando mayores tiempos de extracción, lo que conduce a mayores factores de enriquecimiento. Además, en la HS-SDME se evitan las pérdidas irreproducibles del disolvente debido a su baja presión de vapor y buena estabilidad térmica. Por otro lado, la utilización de ILs como SLM en la HF-LPME ha permitido realizar extracciones estables debido a su alta viscosidad y baja volatilidad, lo que conduce a una buena repetibilidad. Finalmente, su elevada densidad facilita la separación de las fases y sedimentación del extracto en la DLLME. Sin embargo, el hecho más destacable es que los ILs han permitido el desarrollo de nuevas técnicas de microextracción como son la DLLME con control de temperatura (temperature-controlled ionic liquid

DLLME, TC-IL-DLLME) [81] y la DLLME con formación in-situ del IL (in-

situ ionic liquid formation DLLME, in-situ IL-DLLME) [82,83]

En la inmensa mayoría de las publicaciones sobre técnicas de LPME basadas en ILs, se destaca repetidamente el aspecto ecológico de estos disolventes. Sin embargo, aunque los ILs fueron inicialmente propuestos como disolventes alternativos no perjudiciales para el medio ambiente, diversos estudios han demostrado su toxicidad en el medio acuático [84]. Además, se ha


45

comprobado que la síntesis de los ILs ocasiona un mayor impacto medioambiental que la obtención de un disolvente orgánico tradicional [85]. No obstante, para evaluar el impacto medioambiental de un método analítico es importante considerar el procedimiento en su totalidad. Por ejemplo, si se selecciona como técnica de análisis instrumental la GC acoplada a un detector de espectrometría de masas (mass spectrometry, MS), su utilización proporciona un mayor impacto (e.g., consumo energético muy elevado) que el empleo de pequeños volúmenes de IL durante la preparación de la muestra. Además, la baja presión de vapor de los ILs minimiza la contaminación ambiental por evaporación, lo que reduce el peligro de exposición del operador y facilita su reciclaje. Por tanto, desde un punto de vista global, el uso de ILs en técnicas de LPME proporciona más beneficios que perjuicios, aunque se debe mostrar una actitud más crítica a la hora de destacar el aspecto ecológico de un nuevo método o técnica de microextracción que incluya ILs.

Recientemente, se ha presentado una nueva subclase de ILs a los que se ha incorporado en el catión o en el anión un componente paramagnético, normalmente un metal de transición o un lantánido [86]. Estos nuevos líquidos iónicos magnéticos (magnetic ionic liquids, MILs) poseen propiedades físicas y químicas similares a las de los ILs clásicos y además, muestran respuesta frente a un campo magnético externo. La Figura II.26 ilustra los cationes y aniones de los MILs más utilizados hasta la fecha en Química Analítica [86]. Como puede observarse, el metal paramagnético forma parte del anión en todos los casos mostrados. Esto es debido a que los MILs basados en cationes con metales paramagnéticos son más difíciles de sintetizar y menos estables, por lo que su utilización ha sido mucho más limitada [86].


46

Figura II.26. Líquidos iónicos magnéticos utilizados en Química Analítica. Cationes: (a) metilimidazolio; (b) pirrolidinio; (c) piridinio; (d) tetraalquilamonio; (e) tetraalquilfosfonio.

La susceptibilidad magnética de los MILs se ve enormemente afectada por

el tipo de ion paramagnético presente en el anión pero además, por su combinación con diferentes cationes. De forma general, los MILs basados en lantánidos poseen una susceptibilidad magnética superior a la de los basados en metales de transición, sin embargo, el coste de estos materiales también es significativamente mayor. Una alternativa más económica para conseguir MILs de elevada susceptibilidad ha sido la síntesis de disolventes policatiónicos que poseen varios aniones con metales de transición para compensar la carga [86].

Comparando con los ILs, la utilización de los MILs como fase extractante en técnicas de LPME es prácticamente anecdótica. De forma general, los MILs basados en haluros metálicos (Figura II.26) poseen una elevada miscibilidad en agua y cierta inestabilidad frente a la hidrólisis del anión. Por este motivo, los MILs fueron en un primer momento utilizados para el tratamiento de muestras con carácter hidrofóbico [86]. Wang et al. propusieron por primera vez la DLLME basada en MILs para la extracción de herbicidas desde diferentes tipos de aceites vegetales [87,88]. La principal ventaja que aportan los MILs a la DLLME es la eliminación de la etapa de centrifugación, ya que la separación de las fases puede realizarse de forma muy sencilla utilizando un

Líquidos iónicos magnéticos

NN RH3C

N

R2R1

N R4

R1

R3

R2

N

R

P R4

R1

R3

R2

Cationes Aniones

(a) (b) (c)

(d) (e)

FeCl4- FeBr4

- FeCl3Br-

Dy(NO3)4-

CoCl4-

MnCl4- GaCl6

- HoCl6- DyCl6

-

Sm(NO3)4-


47

imán que atraiga a la fase extractante (Figura II.27). Sin embargo, en los trabajos presentados por Wang et al. fue necesaria la adición de partículas ferromagnéticas una vez finalizada la extracción que aportaban la susceptibilidad magnética necesaria para la separación eficiente del MIL utilizando un imán [87,88].

Figura II.27. Esquema general de la microextracción líquido-líquido dispersiva basada en líquidos iónicos magnéticos.

La aplicación de técnicas de LPME basadas en MILs al tratamiento de muestras acuosas es un reto para la comunidad científica y ha despertado recientemente su interés e imaginación. Clark et al. propusieron por primera vez varios MILs hidrofóbicos para la SDME y la DLLME de ADN desde disoluciones acuosas [89]. En la SDME, los autores utilizan un imán en forma de barra para soportar la gota de MIL durante la extracción. En la DLLME, las fases son separadas con la ayuda de un imán, sin la necesidad de añadir ningún reactivo adicional (i.e., partículas ferromagnéticas). Sin embargo, en ambos casos se lleva a cabo el análisis instrumental de la fase acuosa y no del extracto enriquecido una vez finalizada la microextracción [89]. Dos publicaciones posteriores recogen la síntesis de nuevos MILs hidrofóbicos para la DLLME de compuestos aromáticos policíclicos desde muestras de agua [90,91]. Estos MILs están basados en cationes imidazolio, tetraalquilamonio y tretraalquilfosfonio con cadenas carbonadas largas que aportan hidrofobicidad

Adición del MIL y agente dispersante a la muestra

Mezcla dispersada Separación de las fases con un imán


48

al disolvente. No obstante, los autores señalan ciertos problemas experimentales tanto para dispersar la fase extractante como para recogerla tras la extracción, ya que los MILs utilizados quedan pegados en las paredes del vial que contiene la muestra [90]. Además, en una de las publicaciones los analitos son re-extraídos desde el MIL hasta un disolvente orgánico inmiscible (i.e., hexano) [90], mientras que en la otra, el MIL es diluido con acetonitrilo antes de su análisis por LC-UV/vis [91]. Estas etapas adicionales de re-extracción o dilución tienen como posible objetivo la obtención de mejores resultados cromatográficos y evitar el deterioro de las columnas, pero suponen una clara disminución del factor de enriquecimiento. Por último, dos publicaciones recientes sobre MILs en técnicas de LPME se han centrado en su utilización como fase extractante en HS-SDME [92] y, por primera vez, en la extracción de especies inorgánicas (i.e., arsénico) seguida de su determinación mediante espectrometría de absorción atómica en horno de grafito [93].

De la discusión anterior se puede concluir que la utilización de los MILs en técnicas de LPME se encuentra en sus inicios. Los principales retos que se deben afrontar son, por un lado, la búsqueda de MILs hidrofóbicos fáciles de sintetizar y adecuados para el tratamiento de muestras acuosas y, por otro, la compatibilidad de estos disolventes con diferentes técnicas de separación como la GC o la LC.


49

II.1.2.2. Microextracción en gota en espacio de cabeza

En el procedimiento convencional de la HS-SDME, la fase extractante es suspendida de la punta de la aguja de una jeringa situada en el espacio de cabeza por encima de la muestra (Figura II.28). Esta técnica está limitada a la extracción de analitos volátiles o semivolátiles, pero permite la extracción desde matrices sucias o complejas evitando posibles interferencias de las especies no volátiles de la muestra.

Figura II.28. Microextracción en gota en espacio de cabeza.

Además del procedimiento mostrado en la Figura II.28, también es posible realizar extracciones en espacio de cabeza en modo dinámico [94] (Figura II.29). En este modo, el disolvente orgánico permanece en el interior de la jeringa mientras que la fase gaseosa es introducida y expulsada repetidamente de su interior. Así, se consigue evitar uno de los principales inconvenientes del modo estático como es la inestabilidad de la gota suspendida. La HS-SDME en modo dinámico puede realizarse de forma manual [94]. Sin embargo, se han desarrollado métodos semiautomáticos [95]

Jeringa

Muestra Barra magnética agitadora

Espacio de cabeza

Gota


50

y automáticos [96,97] con el fin de reducir la manipulación de la muestra y mejorar la reproducibilidad.

Figura II.29. Esquema del procedimiento de microextracción en espacio de cabeza en modo dinámico: (1) fase extractante en el interior de la jeringa; (2) introducción de la fase gaseosa en la jeringa; (3) expulsión de la fase gaseosa.

La HS-SDME puede ser considerada como un sistema de tres fases ya que los analitos son extraídos desde la muestra (fase acuosa) al espacio de cabeza (fase gaseosa) y después a la gota de disolvente (fase orgánica). La Figura II.30 muestra un esquema del proceso de extracción con las diferentes fases implicadas.

Figura II.30. Esquema del sistema de transporte de analito entre las tres fases implicadas en la microextracción en gota en espacio de cabeza y sus áreas interfaciales (ga= gaseosa/acuosa; og= orgánica/gaseosa). (Fuente: referencia [98]).

Orgánica

Gaseosa

Acuosa

Aog

Aga

(2)

Disolvente orgánico

Émbolo de la jeringa

Fase gaseosa

Fina película de disolvente orgánico

(1)

Disolvente orgánico enriquecido con el analito

(3)


51

El transporte de analito durante la HS-SDME ha sido descrito teóricamente mediante un modelo cinético de transporte [98]. Dicho modelo está fundamentado en las siguientes premisas: (i) el flujo del analito entre dos fases es proporcional a la diferencia de concentraciones entre el seno y la interfase; (ii) el analito no se acumula en las interfases; (iii) siempre se alcanza el estado de equilibrio entre las fases adyacentes; y (iv) la cantidad de analito en el espacio de cabeza es pequeño, comparado con las otras fases, y permanece constante (aproximación del estado estacionario). Los tres primeros supuestos están basados en el modelo cinético propuesto para sistemas de extracción de dos fases. Por otro lado, la aproximación del estado estacionario permite simplificar las ecuaciones matemáticas obtenidas y proporcionar una ecuación cinética manejable [98]. La resolución de dicha ecuación relaciona la concentración del analito en la fase orgánica (Co) con el tiempo, tal y como se muestra en la ecuación (1):

Co(t)=Co∞ (1-exp(-λt)) (1)

donde, Co

∞ es la concentración del analito en la fase orgánica en el equilibrio y λ es la constante de velocidad. Esta constante viene dada por la ecuación (2), considerando que la difusión de los analitos en el espacio de cabeza es más rápida que en las otras fases y que, por tanto, el control de la velocidad del proceso reside en la fase acuosa y/o en la fase orgánica:

λ= AogAgakoka

VoKga(AogkoKog+Aga(ka/Kga))(Koa

Vo

Va+1) (2)

donde,

Aga y Aog son las áreas interfaciales entre la fase gaseosa y la fase acuosa, y entre fase orgánica y la fase gaseosa, respectivamente,

ka y ko son los coeficientes de transferencia de masa en la fase acuosa y en la fase orgánica, respectivamente,

Va y Vo son los volúmenes de la fase acuosa y de la fase orgánica, respectivamente,


52

Kog, Kga y Koa son las constantes de distribución en el equilibrio entre la fase orgánica y la fase gaseosa, entre la fase gaseosa y la fase acuosa, y entre la fase orgánica y la fase acuosa, respectivamente.

Un examen detallado de la ecuación (2) permite conocer si la velocidad del proceso de extracción está más influenciada por la fase acuosa, por la fase orgánica o por ambas. Para analitos volátiles con valores de Kog del orden de 103 y Kga del orden de 10-1, ambos términos del denominador son de magnitud similar, asumiendo que Aog y ko son un orden de magnitud inferior que Aga y ka (lo que resulta lógico teniendo en cuenta la geometría mostrada en la Figura II.30 y la agitación de la muestra típicas de la HS-SDME). Por lo tanto, el modelo propuesto sugiere que tanto la transferencia desde la fase acuosa como la transferencia hacia la fase orgánica son etapas limitantes de la velocidad para la extracción de especies volátiles. Desde un punto de vista práctico, esto significa que tanto la agitación de la muestra como de la fase orgánica (i.e., modo dinámico) favorecen el proceso de extracción. Por otro lado, para analitos semivolátiles con valores muy pequeños de Kga, el segundo término del denominador es el dominante de manera que la ecuación (2) de la constante de velocidad puede simplificarse dando lugar a la ecuación (3):

λ≈ Aogko

Vo(Koa

Vo

Va+1) (3)

De acuerdo con la ecuación (3), el control de la velocidad del proceso de

extracción de especies semivolátiles reside en su difusión en la fase orgánica. Por tanto, la extracción de estas especies se verá favorecida con el modo dinámico de extracción.

El modelo cinético propuesto ha sido comprobado experimentalmente utilizando benceno, tolueno, etilbenceno y o-xileno como analitos modelo [98]. Los resultados experimentales obtenidos concuerdan en gran medida con lo predicho por la teoría pero sugieren que la aproximación del estado estacionario no se cumple en el caso de analitos muy volátiles, ya que su


53

concentración en el espacio de cabeza es bastante importante y no puede ser despreciada.

Teniendo en cuenta el modelo teórico que se acaba de describir, es fácil señalar los principales parámetros experimentales que afectan a la HS-SDME. Por un lado, el disolvente utilizado como fase extractante juega un papel fundamental, debiendo cumplir con las siguientes características generales: (i) buena extractabilidad de los analitos de interés; (ii) baja volatilidad para evitar pérdidas por evaporación; (iii) viscosidad y tensión superficial adecuadas para proporcionar estabilidad a la gota suspendida en la punta de la aguja de la jeringa; y (iv) compatibilidad con la técnica de análisis seleccionada. Tradicionalmente, se han utilizado fases extractantes basadas en disolventes orgánicos como el 1-octanol, el hexadecano, el dodecano o el decano. Sin embargo, estos disolventes presentan un punto de ebullición relativamente bajo y las gotas que forman son bastante inestables. Para superar los inconvenientes señalados, se ha propuesto la utilización de ILs como fases extractantes alternativas [77]. Los ILs poseen una elevada estabilidad térmica y baja volatilidad e inflamabilidad. Además, su viscosidad, tensión superficial y densidad permiten la formación de gotas estables de mayor tamaño (e.g., 10 µL), frente a los 1-3 µL que pueden utilizarse con un disolvente orgánico tradicional. Por otro lado, la agitación de la muestra es fundamental para reducir los tiempos de extracción ya que favorece la difusión de los analitos hacia el espacio de cabeza. En HS-SDME, la ausencia de un contacto directo entre las fases permite agitar de forma vigorosa la muestra sin peligrar la estabilidad de la gota suspendida. En cuanto al volumen del espacio de cabeza, éste debe ser reducido al máximo con el fin de facilitar la acumulación de analitos volátiles en él y su transporte hacia la gota, mejorando así la eficacia de la extracción. Finalmente, la temperatura es otro parámetro importante a considerar ya que afecta tanto a la cinética como a la termodinámica de los procesos de transferencia de masa (i.e., constantes de distribución entre las diferentes fases). Por un lado, el aumento de la temperatura de la muestra facilita la transferencia de los analitos hacia el espacio de cabeza y, por tanto, aumenta su concentración en la fase gaseosa. Sin embargo, la transferencia de los analitos a la fase orgánica es un proceso exotérmico por lo que disminuye al aumentar la temperatura. Por tanto, es necesario encontrar una situación de


54

compromiso en la que la HS-SDME tenga lugar bajo las condiciones óptimas de temperatura. Además de los sistemas tradicionales de calentamiento basados en conducción/convección (i.e., placas calefactoras), se ha utilizado la radiación de microondas [99] y la energía de ultrasonidos [100] para acelerar el calentamiento de la muestra manteniendo la fase extractante a temperatura ambiente.

II.1.2.3. Extracción con electromembrana

La EME está basada en la técnica HF-LPME de tres fases en la que se utiliza una diferencia de potencial entre la muestra y la fase aceptora como fuerza motriz para el transporte de analitos cargados a través de la SLM. La Figura II.31 muestra un esquema de la configuración utilizada en el primer trabajo de EME [71], la cual ha sido adoptada en la mayoría de publicaciones posteriores. Como se observa en dicha figura, un electrodo es introducido en el interior de la fibra hueca y otro electrodo es sumergido en la muestra, mientras que ambos son conectados a una fuente de alimentación externa. La aplicación de una diferencia de potencial entre los dos electrodos provoca la transferencia de analitos cargados desde la muestra hasta la fase aceptora acuosa mediante migración electrocinética. La EME es adecuada para la extracción de analitos ionizables cuya carga puede ser ajustada mediante cambios en el pH. Para la extracción de especies básicas, el ánodo (electrodo positivo) es introducido en la muestra mientras que el cátodo (electrodo negativo) es sumergido en la fase aceptora. Ambas fases son acidificadas a un valor de pH por debajo del pK de los analitos, con el fin de asegurar su carga positiva. Para la extracción de especies ácidas, la dirección del campo aplicado es invertida y tanto la muestra como la fase aceptora son ajustadas a un valor de pH que asegure la ionización negativa de los analitos (i.e., pH˃pK).


55

Figura II.31. Extracción con electromembrana.

Entre las principales ventajas de la EME podemos destacar: (i) la fase aceptora se encuentra protegida en el interior de la fibra por lo que se pueden realizar extracciones desde muestras complejas con matrices sucias obteniendo extractos muy limpios; (ii) la selectividad de la extracción puede ser modulada mediante cambios en la magnitud y en la dirección del potencial aplicado; (iii) se pueden utilizar altas velocidades de agitación al no existir riesgo de desprendimiento de la fase extractante, siempre que se evite la formación de burbujas en la superficie de la membrana; (iv) el área de contacto entre las fases es grande; (v) los tiempos de extracción son más cortos que en la HF-LPME debido a la mejora de la transferencia de masa por acción del potencial aplicado; (vi) la fibra puede ser desechada después de un único uso debido a su bajo precio evitando efectos de memoria entre extracciones; y (vii) la técnica puede ser automatizada.

Los aspectos teóricos y fundamentales de la EME han sido descritos en la bibliografía utilizando diferentes modelos [101–103]. Todos estos modelos están basados en simplificaciones, por tanto, ninguna de las ecuaciones

Fase aceptora

Fibra hueca (SLM)

Fuente de energía

Electrodo Electrodo

+/- -/+

Muestra Barra magnética

agitadora


56

matemáticas propuestas explica completamente el proceso de extracción pero son muy útiles ya que permiten mejorar la comprensión de los diferentes parámetros experimentales implicados. Entre los modelos propuestos podemos diferenciar el modelo del estado estacionario y el modelo transitorio.

i. Modelo del estado estacionario

El modelo del estado estacionario está basado en la ecuación de flujo de Nernst-Planck y tiene como asunciones básicas que la SLM está libre de convección y es eléctricamente neutra en todos sus puntos a nivel microscópico, y que todas las especies del sistema poseen una única carga [101]. El flujo de un analito (J) a través de la SLM es calculado mediante la ecuación (4):

J= -Dh

(1+v

lnχ) (

χ-1χ-exp (-v)

) (ch-c0exp (-v)) (4)

donde, D es el coeficiente de difusión del analito, que viene definido por la

ecuación (5) en la que k es la constante de Boltzmann, T es la temperatura absoluta, η es la viscosidad del disolvente inmovilizado en la SLM, y R es el radio del analito:

D= kT6πηR (5)

h es el grosor de la SLM, v es una variable directamente proporcional a la diferencia de

potencial aplicado, χ es la relación entre la concentración total de iones en la muestra y

en la fase aceptora (i.e., balance iónico), ch y c0 son las concentraciones del analito en la interfase

SLM/muestra y fase aceptora/SLM, respectivamente.

De acuerdo con la ecuación (4), el modelo del estado estacionario sugiere que el flujo a través de la SLM aumenta con la diferencia de potencial aplicado


57

(expresado a través de v) y al disminuir el balance iónico a través de la misma, es decir, cuando la concentración de iones en la fase aceptora es grande comparada con la concentracion de iones de la muestra. Este comportamiento ha sido verificado experimentalmente observándose una gran concordancia entre la predicción teórica y los resultados experimentales obtenidos en el laboratorio [101]. Por otro lado, combinando las ecuaciones (4) y (5), se observa la relación de J con la viscosidad del disolvente orgánico inmovilizado en los poros de la fibra, de manera que la transferencia de masa se ve favorecida con disolventes de baja viscosidad. Además, J es inversamente proporcional al grosor de la SLM, lo que implica que fibras huecas con pequeño grosor de pared mejoran la extracción, siempre que esté asegurada su estabilidad mecánica. Finalmente, los coeficientes de distribución del analito entre las diferentes fases juegan también un papel importante en el proceso de extracción ya que afectan a las concentraciones en las interfases (i.e., ch y c0).

ii. Modelo transitorio

El modelo transitorio es un modelo cinético de transporte que incluye la influencia del tiempo sobre la distribución del analito entre las diferentes fases [102]. El modelo transitorio asume que: (i) existe un tiempo de retraso determinado por la permanencia del analito en la SLM; (ii) la etapa limitante de la velocidad es el transporte a través de la SLM; (iii) existe suficiente convección en la muestra y en la fase aceptora como para que el transporte de masa en ambas fases no limite la velocidad del proceso; y (iv) el transporte a través de la SLM es unidireccional y después de un tiempo suficientemente largo, el analito es completamente recogido en la fase aceptora [102]. De acuerdo con el modelo transitorio, la concentración del analito en la muestra (CD), en la fase aceptora (CA) y en la SLM (CSLM) viene dada en función del tiempo a través de las ecuaciones (6), (7) y (8), respectivamente:

CD(t)=CD0 exp (

-AfPD→A

VDt) (6)


58

CA(t)=VDCD

0 -CD(t)(VD+Kd*VSLM)

VA (7)

CSLM(t)=VD(CD

0 -CD(t))-VACA(t)

VSLM (8)

donde,

VD, VA y VSLM son los volúmenes de la muestra, la fase aceptora y la SLM, respectivamente,

CD0 es la concentración inicial del analito en la muestra,

Af es el área superficial activa de la SLM, PD→A es el coeficiente de permeabilidad de la SLM para el

transporte del analito desde la muestra hacia la fase aceptora, Kd

* es el coeficiente de distribución del analito, que viene definido por la ecuación (9) en la que z es la carga, F es la constante de Faraday, R es la constante de los gases, T es la temperatura absoluta, ∆0

ωφ es la diferencia de potencial de Galvani entre la muestra y la SLM, y ∆0

ωφi0 describe la hidrofobicidad del analito.

Kd∗=exp (

zFRT

(∆0ωφ-∆0

ωφi0)) (9)

Una representación gráfica general de las ecuaciones (6), (7) y (8) permite comprender mejor el proceso de extracción. La Figura II.32 muestra la variación con el tiempo de la concentración del analito en las diferentes fases involucradas propuesta por el modelo transitorio.


59

Figura II.32. Variación de la concentración del analito en la muestra, en la membrana líquida soportada y en la fase aceptora en función del tiempo según el modelo transitorio. (Fuente: referencia [103]).

Como se observa en la Figura II.32, inicialmente existe un incremento rápido de la concentración del analito en la SLM, acompañado de una disminución de la misma en la muestra. Después de un corto periodo de tiempo (tiempo de retraso), la concentración del analito en la fase aceptora comienza a aumentar hasta que el sistema alcanza el estado estacionario en el que no existe transferencia neta de masa y la EME debe darse por finalizada. De acuerdo con la ecuación (6), el proceso de extracción tarda más en alcanzar el estado estacionario al aumentar VD, por lo que desde un punto de vista práctico, es preferible utilizar pequeños volúmenes de muestra para lograr extracciones más rápidas. Por otro lado, un aumento en VSLM afecta principalmente al tiempo de retraso ya que el analito necesita más tiempo para atravesar el disolvente orgánico inmovilizado. Por tanto, un menor volumen de SLM acelera el proceso de extracción, aunque dicho volumen viene normalmente determinado por las dimensiones y porosidad de la fibra hueca. El coeficiente de permeabilidad (PD→A) determina la movilidad electrocinética a través de la SLM del analito cargado y depende de la composición química de la SLM y de las propiedades físico-químicas del analito. Altos coeficientes de permeabilidad conducen a extracciones más rápidas y su optimización está

Muestra (CD) SLM (CSLM) Fase aceptora (CA)

Tiempo (min)

0 2 0

4 6 8

2

4

6

8

Con

cent

raci

ón (m

ol c

m-3

)

10*10-6


60

relacionada con la selección del disolvente orgánico apropiado. Finalmente, el coeficiente de distribución Kd

* relaciona directamente la diferencia de potencial aplicado con la velocidad de la extracción.

La validez del modelo transitorio ha sido comprobada experimentalmente obteniendo resultados que ratifican, de forma general, el modelo propuesto [103]. Sin embargo, se han observado ciertas desviaciones de la teoría que han podido ser fácilmente explicadas. Por ejemplo, algunos trabajos han señalado una disminución en la concentración del analito en la fase aceptora con tiempos largos de extracción. Este hecho puede estar relacionado con la degradación parcial de la SLM o con cambios inesperados en el pH de la fase aceptora debido a las reacciones de electrolisis que se muestran en las ecuaciones (10) y (11) [104,105]:

Cátodo: 2H+ + 2e- H2 (10)

Ánodo: H2O 2H+ + ½O2 + 2e- (11) Durante la extracción de especies básicas cargadas positivamente, la

reducción de protones en el cátodo, electrodo sumergido en la fase aceptora, puede provocar un aumento paulatino del pH en dicha fase. Con tiempos largos de extracción, este hecho puede afectar a la ionización de los analitos y desembocar en la difusión de especies neutras hacia la SLM, disminuyendo las recuperaciones obtenidas. Durante la extracción de especies ácidas cargadas negativamente, ocurriría un proceso análogo debido a una disminución en el pH en la fase aceptora provocado por la reacción de oxidación del agua en el ánodo. Estos cambios de pH pueden ser mitigados mediante la utilización de disoluciones tamponadas. Por otro lado, diferentes autores han observado que un aumento en la diferencia de potencial aplicado no implica necesariamente una mejora en la extracción, lo que se ha relacionado con situaciones en las que la EME deja de estar limitada por la transferencia de masa a través de la SLM [103].


61

Como se acaba de describir, el modelo del estado estacionario y el modelo transitorio están basados en premisas diferentes y proponen modelos matemáticos dispares. Sin embargo, ambos modelos coinciden en señalar la importancia y el efecto de determinados parámetros experimentales (e.g., diferencia de potencial aplicado, composición de la SLM, etc.) sobre la EME. De todos estos parámetros, la utilización de un disolvente orgánico apropiado tiene una especial relevancia. Dicho disolvente debe poseer una buena extractabilidad de los analitos de interés y una baja viscosidad para facilitar su transporte. Además, no conviene olvidar que el disolvente orgánico debe poseer cierto momento dipolar o conductividad que le permita soportar el flujo de corriente del sistema. Por último, el disolvente debe presentar una alta afinidad y capacidad de inmovilización en los poros de la fibra, baja miscibilidad en agua y baja volatilidad para evitar pérdidas por disolución o evaporación. Entre los disolventes más utilizados destacan el 2-nitrofenil octil éter, el 1-isopropil-4-nitrobenceno y el 1-etil-2-nitrobenceno para la extracción de especies básicas, y el 1-octanol y el 1-heptanol para la extracción de especies ácidas.

Al igual que otras técnicas de LPME, la EME ha experimentado numerosas modificaciones desde sus inicios [106]. Por ejemplo, se han propuesto cambios en la generación de la fuerza motriz para el transporte de los analitos cargados. Tradicionalmente, se ha aplicado una diferencia de potencial constante entre la muestra y la fase aceptora. Sin embargo, recientemente se ha propuesto la aplicación de un potencial a pulsos [107] o de una corriente eléctrica constante [108] con los que se han obtenido sistemas de extracción más estables y reproducibles. Por otro lado, se ha presentado la EME de dos fases en la que la fase aceptora es un disolvente orgánico con suficiente conductividad eléctrica (i.e., 1-heptanol) [109]. La importancia de esta modificación reside en la ampliación de la aplicabilidad de la técnica al hacerla compatible con dispositivos de medida basados en la técnica de GC. Sin embargo, entre todas las modificaciones introducidas, la extracción simultánea de especies ácidas y básicas, y la utilización de dispositivos en formato miniaturizado son las más significativas. La extracción simultánea de especies ácidas y básicas se ha conseguido mediante la utilización de membranas compartimentadas [110] o de dos fibras huecas separadas [111].


62

Esta última configuración, mostrada en la Figura II.33, es más sencilla y robusta, y ha encontrado una mayor aceptación entre los investigadores. Las especies básicas, cargadas positivamente, migran hacia el cátodo sumergido en una fase aceptora acidificada mientras que las especias ácidas, cargadas negativamente, migran hacia el ánodo sumergido en una fase aceptora basificada. El pH de la muestra debe ajustarse a un valor neutro para el que todas las especies se encuentren ionizadas. De este modo, los analitos básicos y ácidos son extraídos simultáneamente pero por separado, por lo que es necesario el mezclado posterior de las fases aceptoras para la determinación conjunta de ambas especies.

Figura II.33. Extracción con electromembrana de especies ácidas y básicas de forma simultánea.

La miniaturización de la EME ha logrado un importante grado de desarrollo, llegando a dispositivos en formato de chip y de nano-EME [112–114]. Todos los formatos miniaturizados han utilizado membranas planas, en lugar de fibras huecas, como soporte para la SLM. En el primer trabajo presentado, los analitos son extraídos desde 10 µL de muestra a 10 µL de fase aceptora (Figura II.34) [112].

Muestra

Fase aceptora básica

Fibra hueca (SLM)

Fuente de energía

Cátodo -

Ánodo +

Fase aceptora ácida


63

Figura II.34. Sistema miniaturizado de extracción con electromembrana. (Fuente: referencia [112]).

Con la EME en formato de chip, se utilizan 25-75 µL de muestra bombeada en flujo continuo y 7 µL de fase aceptora que permanece estancada durante la extracción [113]. Sin embrago, con la nano-EME se ha conseguido reducir el volumen de la fase aceptora hasta el orden de los nL, lo que resulta altamente ventajoso para lograr altos factores de enriquecimiento [114].

II.1.2.4. Microextracción líquido-líquido dispersiva

La DLLME está basada en la dispersión de la fase extractante en el seno de la muestra con el fin de aumentar el área de contacto entre ambas fases y favorecer así la transferencia de masa. En el procedimiento convencional [72], la DLLME incluye un sistema de tres componentes en el que se utiliza un disolvente orgánico más denso que el agua como fase extractante (20-100 µL) y otro disolvente orgánico como agente dispersante (0.1-2 mL). Este último debe ser parcialmente miscible tanto en la fase extractante como en la muestra con el fin de lograr una dispersión homogénea. Tras la extracción, las fases son separadas por centrifugación, y el extracto enriquecido con el analito es recogido para ser analizado. La Figura II.35 muestra un esquema del procedimiento convencional de la DLLME, presentado en el primer trabajo sobre esta técnica [72].

15 V

Muestra (10 µL)

Fase aceptora (10 µL)

Electrodo de platino

Membrana plana (SLM)

Papel de aluminio


64

Figura II.35. Esquema del procedimiento convencional de la microextracción líquido-líquido dispersiva.

Las principales ventajas que presenta la DLLME son: (i) tiempos cortos de extracción debido a la elevada superficie de contacto entre las finas gotas de fase extractante y la muestra; (ii) altos factores de enriquecimiento; y (iii) fácil manejo y eliminación de ciertos inconvenientes relacionados con otras técnicas de LPME, como el desprendimiento de la gota en SDME o la manipulación de la fibra en HF-LPME.

Entre los parámetros experimentales que afectan a la DLLME, los disolventes utilizados como fase extractante y agente dispersante juegan un papel fundamental. El disolvente extractante debe ser más denso que el agua, inmiscible en ella, compatible con la técnica de análisis seleccionada y poseer una buena extractabilidad de los analitos de interés. Tradicionalmente, se han utilizado disolventes halogenados como el clorobenceno, el cloroformo o el tetracloruro de carbono. En cuanto al agente dispersante, es esencial que éste

Retirada de la fase extractante

enriquecida

Inyección de la fase extractante y agente

dispersante en la muestra

Dispersión de la fase extractante

Agitación manual Centrifugación

Fases separadas

Eliminación de la fase

acuosa


65

posea cierta miscibilidad tanto en la fase extractante como en la muestra, siendo el metanol, el etanol, la acetona o el acetonitrilo los disolventes más comunes debido a su bajo coste y toxicidad. Por otro lado, el volumen de agente dispersante también afecta de forma muy significativa a la eficacia de la extracción. Volúmenes demasiado pequeños pueden provocar una mala dispersión de la fase extractante, mientras que volúmenes demasiado grandes pueden aumentar la solubilidad de los analitos en la fase acuosa y reducir su transferencia a la fase orgánica. Además de los parámetros mencionados, el volumen de muestra, su pH y concentración salina, el tiempo de extracción y las condiciones de centrifugación (i.e., velocidad y tiempo) son comúnmente optimizados en las diferentes aplicaciones de la técnica. Sin embargo, conviene enfatizar que el tiempo de extracción no juega un papel crítico a la hora de alcanzar altos factores de enriquecimiento ya que el estado de equilibrio se alcanza de forma casi instantánea debido a la eleva superficie de contacto entre las fases. Como consecuencia, la DLLME puede darse por terminada en unos pocos minutos o incluso segundos, siendo ésta una de las principales virtudes de la técnica.

A pesar de las importantes ventajas que supuso la aparición de la DLLME, su procedimiento convencional presenta algunos inconvenientes, entre los que cabe destacar: (i) la utilización de disolventes orgánicos clorados, volátiles y tóxicos como fase extractante; (ii) la dispersión asistida por un agente dispersante que puede aumentar la solubilidad del analito en la muestra; y (iii) la centrifugación para separar las fases, que muchas veces supone la etapa más lenta del proceso y, además, dificulta su automatización. Con el fin de superar los inconvenientes mencionados, la DLLME ha sido objeto de numerosas modificaciones que han afectado, principalmente, a la utilización de fases extractantes alternativas y de diferentes estrategias para generar la dispersión o separar las fases (Figura II.36).


66

Figura II.36. Modificaciones más significativas de la microextracción líquido-líquido dispersiva.

Junto con los ILs (discutidos anteriormente en el Apartado II.1.2.1.1), los disolventes orgánicos de baja densidad, los surfactantes y las disoluciones acuosas han sido presentados como fases extractantes alternativas a los disolventes orgánicos tradicionales, con el objetivo de desarrollar procedimientos menos perjudiciales para el medio ambiente y la salud del operador. Por un lado, los disolventes orgánicos de baja densidad (i.e., no clorados) son considerados menos tóxicos que los disolventes clorados tradicionales. Sin embargo, su empleo en la DLLME ha obligado a desarrollar nuevas estrategias para recoger la gota de extractante tras la separación de las fases, ya que ésta no sedimenta en el fondo del tubo de extracción sino que queda flotando sobre la superficie de la muestra. Una de las estrategias propuestas está basada en la combinación de la DLLME con la técnica de solidificación de la gota orgánica flotante (solidification of the floating organic

drop, SFOD) [115,116]. En dicha combinación, el tubo de extracción es introducido en un baño de hielo tras la etapa de centrifugación. Seguidamente, el disolvente orgánico solidificado es transferido a un vial vacío utilizando una

• Disolventes orgánicos de baja densidad • Líquido iónicos • Surfactantes • Disoluciones acuosas

Fases extractantes alternativas

• Surfactantes como agente dispersante • Energía de ultrasonidos • Agitación con vórtex • Extracción asistida con aire

Estrategias para generar la dispersión

• Separación espontánea de las fases • Agente desemulsionante

Estrategias para separar las fases


67

pequeña espátula y, una vez allí, funde rápidamente a temperatura ambiente para su posterior análisis. El procedimiento descrito requiere de disolventes orgánicos que posean un punto de fusión cercano a la temperatura ambiente (entre 10 y 30 ºC), además de una baja solubilidad en agua, baja volatilidad y menor densidad que la fase acuosa [117]. El número limitado de disolventes adecuados para la DLLME-SFOD ha provocado que algunos autores se decanten por la utilización de tubos de extracción especiales. Estos tubos cuentan con un cuello estrecho en el que se acumula la fase extractante menos densa que el agua facilitando así su retirada (Figura II.37) [117].

Figura II.37. Ejemplos de tubos de extracción especiales diseñados para la microextracción líquido-líquido dispersiva con disolventes de baja densidad. (Fuente: referencia [117]).

Por otro lado, los surfactantes son sustancias anfipáticas que pueden

agregarse dando lugar a disolventes supramoleculares, también llamados coacervatos, con diferentes configuraciones (e.g., micelas acuosas, micelas reversas, vesículas). La principal ventaja que presentan estos disolventes es la gran variedad de interacciones que pueden tener con el analito, al contar con regiones de diferente polaridad. Además, estas interacciones pueden ser modificadas variando los grupos polares y apolares de los surfactantes por lo que en cierta medida, se puede diseñar el disolvente supramolecular más adecuado para una aplicación específica [118,119]. Los disolventes supramoleculares poseen una densidad menor que la del agua, por lo que también ha sido necesario el desarrollo de dispositivos de extracción especiales (similares a los mostrados en la Figura II.37) para facilitar su retirada una vez


68

finalizada la extracción. Finalmente, las disoluciones acuosas también son consideradas más respetuosas con el medio ambiente que los disolventes orgánicos tradicionales o los ILs. Sin embargo, como principal limitación podríamos mencionar que su utilización está restringida a la extracción de especies hidrofílicas desde muestras con carácter hidrofóbico (e.g., aceites).

En cuanto a las estrategias para generar la dispersión, la DLLME ha empleado tradicionalmente disolventes orgánicos polares (e.g., metanol, acetona, acetonitrilo) para favorecer la dispersión de la fase extractante en la muestra. Alternativamente, se ha propuesto la utilización de surfactantes (e.g., dodecilsulfato de sodio, Triton-X, Aliquat 336), ya que al tratarse de especies anfipáticas disminuyen la tensión superficial entre las fases facilitando su emulsión [120]. Comparando con los disolventes orgánicos tradicionales, son necesarias menores cantidades de surfactante para conseguir una dispersión adecuada y, además, se trata de especies menos tóxicas y contaminantes. Por otro lado, siguiendo la tendencia de reducir el consumo de reactivos y desarrollar procedimientos más sostenibles, se han propuesto varias estrategias para generar la dispersión con las que se evita la utilización de cualquier agente dispersante. Entre las más destacadas se encuentran la utilización de la energía de ultrasonidos y la agitación con vórtex [121].

Finalmente, la etapa de centrifugación es considerada la más lenta y uno de los puntos débiles de la DLLME, ya que dificulta enormemente su automatización. Como consecuencia, algunos autores han trabajado en el desarrollo de estrategias alternativas que permiten abrir nuevos horizontes hacia la automatización de la técnica. Una de estas estrategias aprovecha la tendencia espontánea de las fases a separarse una vez finalizada la agitación de la mezcla. Con este procedimiento, se han utilizado tanto disolventes de elevada densidad, que se depositan en el fondo del tubo de extracción (e.g., IL, tetracloruro de carbono) [122,123], como disolventes de baja densidad que quedan flotando sobre la superficie de la muestra [124]. Otra de las alternativas propuestas está basada en la incorporación de un volumen adicional de disolvente orgánico (normalmente el mismo que el utilizado como agente dispersante) con el fin de romper la emulsión de las fases [125,126]. De este modo, se reduce drásticamente el tiempo global del procedimiento ya que la


69

separación de las fases ocurre de forma instantánea. Sin embargo, como principal inconveniente podemos destacar la utilización de grandes cantidades de disolvente orgánico volátil y tóxico que, además, puede favorecer la disolución del analito en la fase acuosa reduciendo la eficacia de la extracción.

Las diferentes modificaciones mostradas en la Figura II.36 han dado lugar

a nuevas modalidades de DLLME recogidas en excelentes revisiones sobre la evolución de esta técnica [127–133]. De todas estas modalidades, a continuación se describirán en apartados individuales aquéllas que han sido objeto de estudio en el trabajo de investigación que recoge la presente memoria.

II.1.2.4.1. Microextracción líquido-líquido dispersiva con formación in-situ del líquido iónico

En el año 2009, Baghdadi y Shemirani [82], y Jao y Anderson [83] propusieron de manera prácticamente simultánea la técnica in-situ IL-DLLME. Aunque los primeros autores llamaron a esta nueva técnica “microextracción con formación in-situ del disolvente”, el nombre propuesto por los segundos (i.e., in-situ IL-DLLME) ha sido adoptado con más frecuencia en publicaciones posteriores. La in-situ IL-DLLME está basada en la formación de la fase extractante en el seno de la muestra mediante una reacción de intercambio iónico entre un IL hidrofílico y una sal. De acuerdo con el procedimiento general, el IL hidrofílico es disuelto en la fase acuosa y, seguidamente, se añade la sal dando lugar a una reacción de metátesis con la que se forman pequeñas gotas uniformemente dispersas de un IL hidrofóbico que actúa como fase extractante (Figura II.38).

Figura II.38. Reacción de metátesis durante la microextracción líquido-líquido dispersiva con formación in-situ del líquido iónico.

Cl- + KPF6 PF6

- + KCl

Líquido iónico hidrofílico Líquido iónico hidrofóbico


70

La reacción de metátesis y la extracción de los analitos tienen lugar en un único paso en el que el contacto de las fases es máximo, haciendo que la transferencia de masa sea muy rápida y eficaz. Además, otra de las principales ventajas de esta técnica consiste en la obtención del IL hidrofóbico dispersado, evitando así la utilización de agentes dispersantes o de dispositivos adicionales para generar la dispersión (e.g., baño de ultrasonidos o agitador tipo vórtex). Sin embargo, en alguna ocasión se ha recurrido a la energía de ultrasonidos, la radiación de microondas o la agitación con vórtex para mejorar la cinética de la reacción de intercambio iónico [134].

Los ILs hidrofílicos más utilizados en la in-situ IL-DLLME están basados en cationes imidazolio con cadenas carbonadas cortas (e.g., 1-butil-3-metilimidazolio, 1-hexil-3-metilimidazolio, 1-octil-3-metilimidazolio) y aniones cloruro o tetrafluoroborato. Sin embargo, recientemente se ha propuesto la utilización de ILs hidrofílicos que contienen cationes imidazolio con cadenas carbonadas de mayor longitud (e.g., 1-hexadecil-3-butilimidazolio). En cuanto a las sales de intercambio iónico, las más comunes son el KPF6, el NaPF6 o el bis[(trifluorometil)sulfonil]imida de litio [134].

II.1.2.4.2. Microextracción líquido-líquido dispersiva en fase reversa

El empleo de disoluciones acuosas como fase extractante tiene como objetivo la extracción de especies hidrofílicas desde muestras con carácter hidrofóbico (e.g., aceites vegetales, extractos orgánicos, biodiesel, etc.). Este nuevo concepto de microextracción en el que la naturaleza de la fase extractante y la muestra se invierte, ha dado lugar a la técnica de microextracción líquido-líquido dispersiva en fase reversa (reversed-phase

DLLME, RP-DLLME). La RP-DLLME fue presentada por Hashemi et al. en el año 2010 [135]. Aunque en este primer trabajo se utilizó agua como fase extractante, publicaciones posteriores han empleado disoluciones acuosas tamponadas a diferentes pH y mezclas agua/metanol o agua/etanol en diferentes proporciones para la extracción de los analitos. Por otro lado, la dispersión de la fase extractante ha sido asistida con frecuencia por disolventes orgánicos como el acetato de etilo o el 1,4-dioxano [135–138]. Además, en


71

algunos trabajos también se ha utilizado la energía de ultrasonidos [138,139] o la agitación con vórtex [137] para mejorar la formación de la emulsión.

La RP-DLLME se ha revelado como una técnica simple y rápida para el tratamiento de muestras con carácter hidrofóbico, principalmente aceites vegetales comestibles, logrando altos factores de enriquecimiento y extractos muy limpios. Además, otra de sus principales ventajas es la eliminación del disolvente orgánico clorado del procedimiento. Sin embargo, las muestras de aceite son normalmente diluidas antes de la microextracción con volúmenes relativamente elevados (2-5 mL) de disolventes orgánicos como el ciclohexano o el hexano, lo que podría considerarse el punto débil de la técnica. No obstante, dichos volúmenes siguen siendo despreciables comparados con los consumidos en otras técnicas de extracción como la SPE o la LLE, tradicionalmente utilizadas para el tratamiento de este tipo de muestras.

II.1.2.4.3. Microextracción líquido-líquido dispersiva asistida por energía de ultrasonidos o por agitación con vórtex

Los ultrasonidos son ondas mecánicas de frecuencia superior a los 20 kHz que se propagan en un medio elástico generando zonas alternas de compresión y expansión (i.e., rarefracción). En estas regiones de presión cambiante se produce el fenómeno de cavitación. Este fenómeno está basado en la formación de pequeñas burbujas de gas que se expanden durante la rarefracción y que se contraen durante la compresión hasta alcanzar un punto crítico en el que la energía de ultrasonidos no es suficiente para retener la fase vapor en la burbuja y colapsan (i.e., implosionan) generando regiones localizadas de elevada presión y temperatura que alcanzan los 50 MPa y los 5500 ºC (Figura II.39) [140]. En un sistema heterogéneo en el que existen dos fases líquidas inmiscibles, la principal consecuencia de la radiación por ultrasonidos consiste en la fragmentación de una de las fases dando lugar a emulsiones con gotas de tamaño muy pequeño. Además, cuando las burbujas de cavitación colapsan cerca de la interfase líquido-líquido, los incrementos momentáneos y localizados de presión y temperatura pueden afectar al intercambio de componentes entre ambas fases [141].


72

Figura II.39. Fenómeno de cavitación producido por las ondas de ultrasonidos.

Teniendo en cuenta los fenómenos descritos, Regueiro et al. propusieron la DLLME asistida por ultrasonidos (ultrasound-assisted DLLME, UA-DLLME) en el año 2008 [141]. En este primer trabajo se utilizó la energía de ultrasonidos para generar la dispersión de la fase extractante obviando el agente dispersante. Sin embargo, numerosas publicaciones posteriores han incluido un disolvente orgánico o un surfactante con el fin de favorecer la formación de la emulsión. En las diferentes aplicaciones de la UA-DLLME, se han utilizado dispositivos muy diversos para la radiación de los ultrasonidos, que van desde los típicos baños y sondas hasta los sonorreactores [142]. Los baños de ultrasonidos son los más populares debido a su bajo coste y fácil disponibilidad. Sin embargo, estos dispositivos presentan algunos inconvenientes como por ejemplo, la falta de uniformidad de la energía suministrada y la disminución de la potencia con el tiempo, lo que puede provocar problemas de repetibilidad entre extracciones. Además, los baños de ultrasonidos aplican la energía de manera indirecta ya que ésta debe atravesar las paredes del tubo de extracción que contiene la muestra, además del agua que se utiliza para transmitir la energía desde el transductor hasta el tubo. En los sonorreactores, la energía también es suministrada de forma indirecta aunque su potencia es 50 veces superior a la de los baños de ultrasonidos. Por el contrario, las sondas de ultrasonidos (i.e., sonotrodos) son sumergidas

compresión presión máxima

presión mínima

rarefracción

formación de la burbuja

crecimiento implosión

50 MPa 5500 ºC


73

directamente en la muestra, proporcionando una energía mucho más uniforme y eficiente [143].

A pesar de las ventajas inherentes a la UA-DLLME (e.g., eliminación del agente dispersante, mejoras en la eficacia y velocidad de la extracción), la energía de ultrasonidos es a la vez fuente de ciertos inconvenientes. Por un lado, la presencia de zonas localizadas con condiciones extremas de presión y temperatura, así como la formación de radicales libres pueden provocar la degradación de determinados analitos. Además, los ultrasonidos generan emulsiones altamente estables que son difíciles de romper, por lo que se requieren tiempos largos de centrifugación para la separación de las fases [121]. Estos inconvenientes han estimulado la búsqueda de nuevas estrategias para generar la dispersión entre las que se encuentra la agitación con vórtex. La DLLME asistida por vórtex (vortex-assisted DLLME, VA-DLLME) fue presentada por Yiantzi et al. en el año 2010 [144]. Esta técnica emplea una agitación circular y vibratoria muy vigorosa para lograr la dispersión de la fase extactante en la muestra. Con este tipo de agitación, la emulsión es mucho menos estable que la obtenida con la UA-DLLME y se evita la formación de radicales libres y de zonas de alta presión y temperatura. Además, otro aspecto a favor de la VA-DLLME es que los agitadores tipo vórtex son más económicos y más eficientes energéticamente hablando que los sonorreactores, los baños de ultrasonidos o los sonotrodos [121]. Finalmente, otra ventaja de la VA-DLLME consiste en la eliminación del agente dispersante aunque, al igual que sucede con la UA-DLLME, algunos autores lo siguen incluyendo en el procedimiento.

Diseño estadístico de experimentos

74

II.1.3. Diseño estadístico de experimentos

El diseño estadístico de experimentos, también denominado diseño experimental, es una metodología basada en herramientas matemáticas y estadísticas que tiene como principales objetivos: (i) seleccionar la estrategia experimental óptima que permita obtener la información buscada con el mínimo coste; y (ii) evaluar los resultados experimentales garantizando la máxima fiabilidad de las conclusiones que se obtienen. Por tanto, dicha metodología trata dos aspectos fundamentales: el diseño de los experimentos y el análisis estadístico de los resultados. Ambos aspectos están íntimamente relacionados ya que el método de análisis depende directamente del diseño planteado [145].

Las situaciones en las se puede aplicar el diseño estadístico de experimentos son muy numerosas y variadas. De forma general, se aplica a sistemas o procesos como el mostrado esquemáticamente en la Figura II.40, en los cuales se obtiene una o más respuestas (variables dependientes) que dependen de los valores de una o más variables controlables (variables independientes), también llamadas factores. Además, las respuestas pueden estar afectadas por otras variables no controladas por el experimentador.

Figura II.40. Esquema general de un sistema o proceso en estudio mediante diseño estadístico de experimentos. (Fuente: referencia [145]).

La optimización de un proceso mediante diseño estadístico de experimentos resulta una alternativa muy atractiva e interesante a la

x1

xn … …

y1

yk Proceso

variables no controladas

variables o factores respuestas


75

optimización univariante clásica. Con esta última metodología, también llamada optimización paso a paso o de un-factor-cada-vez, se parte de unas condiciones iniciales para cada factor y se realizan experimentos en los cuales todos los factores se mantienen constantes a un determinado valor o nivel excepto el que se está estudiando. De este modo, la variación de la respuesta se puede atribuir únicamente a la variación del factor en estudio y, por tanto, revela su efecto. Una vez que se ha obtenido la mejor respuesta, el factor estudiado es fijado en su valor óptimo y se repite el procedimiento para otro factor, y así sucesivamente para todos los factores estudiados. La principal desventaja de esta metodología es que no considera la posible interacción entre factores. Existe una interacción cuando uno de los factores no produce el mismo efecto en la respuesta con valores diferentes de otro factor. Las interacciones entre factores son muy comunes y, en caso de existir, el método clásico casi siempre produciría resultados deficientes [145]. Por el contrario, el diseño estadístico de experimentos, en los que se varía de forma simultánea más de un factor, aporta una solución a este problema ya que es capaz de mostrar la existencia de interacciones. Además, la variación simultánea de más de un factor permite obtener información valiosa con un número reducido de experimentos, lo que supone un ahorro de tiempo y recursos.

Para llevar a cabo un diseño estadístico de experimentos de forma satisfactoria es recomendable seguir los pasos o etapas que se enumeran y describen a continuación [145]:

i. Planteamiento del problema y definición del objetivo

Un planteamiento claro del problema a abordar y de los objetivos a cumplir contribuye sustancialmente a alcanzar una mejor compresión de la solución final obtenida. Para ello, es importante recabar toda la información posible acerca del sistema o proceso a estudiar y que pueda ser relevante para los experimentos que se van a realizar.

ii. Selección de los factores y niveles

En esta etapa se deben elegir los factores que se van a estudiar durante el diseño, el rango en el que se hará variar cada factor y los valores o niveles específicos a los que se realizarán los distintos experimentos. Para ello, se


76

requiere un conocimiento profundo del proceso en estudio, el cual puede obtenerse mediante experiencia práctica y/o conocimientos teóricos. Cuando el objetivo del experimento es el cribado de factores (i.e., encontrar qué factores tienen un efecto significativo en la respuesta y cuáles no), se recomienda incluir un número reducido de niveles. En general, se utilizan dos niveles por factor en los estudios de cribado, mientras que el rango en el que se sitúan debe mantenerse lo más amplio posible.

iii. Selección de la variable respuesta

La variable respuesta seleccionada debe proporcionar información útil sobre el proceso que se está estudiando y para la solución del problema planteado. iv. Elección del diseño de experimentos

Para la elección del diseño es importante tener en cuenta los objetivos experimentales (e.g., cribado, optimización). Si las etapas anteriores se han realizado correctamente, esta etapa es relativamente sencilla ya que existen numerosos programas estadísticos que ayudan a abordarla. En dichos programas, se introduce la información sobre el número de factores y niveles, obteniendo una colección de diseños entre los que se debe elegir el más adecuado según las necesidades y objetivos particulares.

v. Realización de los experimentos

Una vez elegido el diseño, el programa estadístico utilizado proporciona una hoja de trabajo (i.e., matriz de experimentos) que incluye los diferentes experimentos y el orden aleatorio en el que deben realizarse. La ejecución en orden aleatorio tiene como objetivo evitar las variables no controlables que pueden ocasionar algún tipo de tendencia o sesgo en la respuesta. Esta etapa debe realizarse con especial cuidado y conforme al diseño planteado, ya que los errores en el procedimiento experimental afectarán a la validez de los resultados y a las conclusiones obtenidas.

vi. Análisis de los datos

Las respuestas obtenidas en cada uno de los experimentos que forman el diseño son introducidas en el programa estadístico que calcula los efectos de


77

los diferentes factores, sus interacciones, significación estadística, etc. Los métodos gráficos suelen ser útiles a la hora de presentar los resultados y analizarlos.

vii. Conclusiones

Una vez analizados los datos, resta sacar conclusiones prácticas de los resultados y tomar decisiones. La ventaja principal de los métodos estadísticos es que agregan objetividad al proceso de toma de decisiones. Por último, es conveniente realizar experimentos que confirmen y validen las conclusiones obtenidas.

El diseño estadístico de experimentos ha sido utilizado para optimizar la etapa de preparación de la muestra en varios de los trabajos presentados en esta memoria. En todos ellos, el objetivo principal es encontrar las condiciones experimentales con las que se obtiene la mayor señal del analito, por lo que dicha señal es utilizada como variable respuesta. Los factores que pueden jugar un papel importante en los diferentes procesos son estudiados en una primera etapa de cribado. Posteriormente, los factores significativos son optimizados para hallar las mejores condiciones en las que la correspondiente preparación de la muestra tiene lugar (i.e., se obtiene una mayor respuesta del analito).

Aunque existen varios tipos de diseños dependiendo de la aplicación específica y de los objetivos a alcanzar, a continuación se describirán los dos tipos de diseño utilizados en esta memoria.

II.1.3.1. Diseño de Plackett-Burman

Los diseños factoriales resultan muy útiles en las etapas iniciales de un trabajo experimental cuando se requiere el estudio de varios factores. Estos diseños incluyen todas las combinaciones posibles entre los diferentes niveles de los factores investigados. Los diseños factoriales de dos niveles (+1,-1) son los que necesitan el menor número de experimentos (2k) para estudiar k factores en un diseño factorial completo. Sin embargo, cuando el número de factores se incrementa, los experimentos necesarios para realizar este tipo de


78

diseño se disparan y, en muchos casos, superan el tiempo y los recursos disponibles. Por ejemplo, para el estudio de 5 factores haría falta realizar 25=32 experimentos. Por este motivo, si es posible suponer que ciertas interacciones de orden superior (entre tres o más factores) son insignificantes, se puede obtener información sobre los efectos principales y las interacciones de orden inferior (entre dos factores) realizando un diseño factorial fraccionado, en los que no se incluyen todas las combinaciones entre niveles [145].

El diseño de Plackett-Burman es un tipo de diseño factorial fraccionado de dos niveles utilizado para estudiar hasta k=N-1 factores con N experimentos, siendo N múltiplo de 4. Si k=N-1, se dice que el diseño está saturado y carece de grados de libertad disponibles para la estimación del error. Este diseño considera que las interacciones entre factores pueden ser totalmente ignoradas por lo que, desde un punto de vista práctico, resulta muy ventajoso ya que permite calcular los efectos principales con un número reducido de experimentos, ahorrando tiempo y recursos [145].

El diseño de Plackett-Burman ha sido utilizado en esta memoria durante las etapas de cribado, cuando inicialmente se considera un número elevado factores que pueden afectar a la respuesta del sistema y se quiere saber cuáles lo hacen de manera significativa (deben ser estudiados en una etapa posterior de optimización) y cuáles no (pueden ser fijados).

La Tabla II.1 muestra a modo de ejemplo una de las matrices de experimentos utilizadas en esta memoria para estudiar k=11 factores en N=12 experimentos. Al considerar dos niveles para cada factor (+1,-1), se supone que la respuesta es aproximadamente lineal entre dichos niveles.


79

Tabla II.1. Matriz de experimentos del diseño de Plackett-Burman para estudiar k=11 factores en N=12 experimentos.

Factores Experimento A B C D E F G H I J K

1 +1 +1 -1 +1 +1 +1 -1 -1 -1 +1 -1 2 -1 +1 +1 -1 +1 +1 +1 -1 -1 -1 +1 3 +1 -1 +1 +1 -1 +1 +1 +1 -1 -1 -1 4 -1 +1 -1 +1 +1 -1 +1 +1 +1 -1 -1 5 -1 -1 +1 -1 +1 +1 -1 +1 +1 +1 -1 6 -1 -1 -1 +1 -1 +1 +1 -1 +1 +1 +1 7 +1 -1 -1 -1 +1 -1 +1 +1 -1 +1 +1 8 +1 +1 -1 -1 -1 +1 -1 +1 +1 -1 +1 9 +1 +1 +1 -1 -1 -1 +1 -1 +1 +1 -1 10 -1 +1 +1 +1 -1 -1 -1 +1 -1 +1 +1 11 +1 -1 +1 +1 +1 -1 -1 -1 +1 -1 +1 12 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1

En alguno de los diseños de Plackett-Burman realizados, se han incluido factores ficticios o factores fantasma. El cambio de nivel de estos factores no supone ningún cambio físico o real en la experimentación y sus efectos son utilizados para estimar el error experimental. Si el efecto de un factor que en realidad no existe resulta significativo, hay algún problema experimental que debe corregirse. En un diseño de Plackett-Burman saturado con N=12 experimentos, tendremos un determinado número de factores reales y hasta N-1 factores fantasma [146].

II.1.3.2. Diseño central compuesto

El diseño central compuesto circunscrito o diseño central compuesto (central composite design, CCD) es aquel cuya superficie de respuesta se ajusta a un polinomio de segundo grado como los mostrados en las ecuaciones (12) y (13), según se estén estudiando 2 o 3 factores, respectivamente. Este diseño es utilizado para evaluar los efectos principales, las interacciones entre factores y los efectos cuadráticos de los factores considerados.


80

y=β0+β1x1+β2x2+β12x1x2+β11x1

2+β22x22 (12)

y=β0+β1x1+β2x2+β3x3+β12x1x2+β13x1x3+β23x2x3+β11x1

2+β22x22+β33x3

2 (13)

donde,

y es la variable respuesta, x1, x2 y x3 son los factores independientes, β0 es una constante y β1-β33 son los coeficientes.

El CCD está basado en un diseño factorial de dos niveles (2k) aumentado con un punto central que se repite n veces y con 2k puntos estrella. El punto central de cada factor es cero y el diseño es simétrico en torno a él. Los puntos estrellas son situados a una distancia axial +α y –α del punto central. Dicha

distancia es fijada a un valor α=√2k4 con el fin de asegurar la rotabilidad del

modelo, lo cual implica que la varianza es la misma para todos los puntos situados a igual distancia del centro del dominio. El punto central es repetido n veces para asegurar la ortogonalidad, la cual garantiza que el efecto de un factor o interacción pueda estimarse de manera independiente del efecto de cualquier otro factor o interacción presente en el modelo [145].

El CCD es muy utilizado en la optimización de procesos. En la presente memoria, se ha utilizado para hallar el valor óptimo de los factores cuyos efectos sobre la respuesta del sistema resultaron significativos durante la etapa previa de cribado (i.e., diseño de Plackett-Burman). La Tabla II.2 muestra a modo de ejemplo la matriz de experimentos de un CCD utilizado para la optimización de k=2 factores. Como se puede observar, los puntos estrellas

están situados a una distancia axial α=√224=1.41, mientras que el experimento

correspondiente al punto central (0,0) se repite un total de 5 veces.


81

Tabla II.2. Matriz de experimentos de un diseño central compuesto para el estudio de k=2 factores en N=13 experimentos.

Factores Experimento A B

1 -1 -1 2 +1 -1 3 -1 +1 4 +1 +1 5 -1.41 0 6 1.41 0 7 0 -1.41 8 0 1.41 9 0 0 10 0 0 11 0 0 12 0 0 13 0 0

Sistemas miniaturizados de detección

82

II.2. Sistemas miniaturizados de detección Cuando se habla de la miniaturización de un método de análisis conviene

considerar por separado dos partes importantes del proceso analítico total: (a) la preparación de la muestra; y (b) el sistema de separación/detección. La miniaturización de la preparación de la muestra, concretamente de las técnicas de extracción, ha logrado un gran desarrollo en los últimos años dando lugar a las diferentes técnicas de microextracción en fase sólida y en fase líquida que se han descrito en esta memoria. Sin embargo, la miniaturización de los sistemas de separación/detección no ha alcanzado un desarrollo tan amplio debido, principalmente, a la complejidad de la instrumentación analítica y de la tecnología necesaria para reducir su tamaño. Los progresos recientes en microfabricación y microfluídica, así como la utilización de nanomateriales, han permitido avanzar en el desarrollo de sistemas dinámicos de separación (e.g., GC y LC) y detectores (e.g., MS) de reducidas dimensiones [147–149]. No obstante, estos sistemas no se encuentran al alcance de muchos laboratorios y siguen siendo escasamente utilizados en la actualidad.

Por otro lado, el acoplamiento de técnicas de LPME con sistemas de detección existentes en formato miniaturizado es muy escaso en la literatura. El grupo de investigación del Profesor Valcárcel ha propuesto la utilización de la espectrometría de movilidad iónica (ion mobility spectrometry, IMS) acoplada a la técnica de HS-SDME basada en ILs [150–152]. La IMS caracteriza a las moléculas por la movilidad en fase gaseosa de sus iones formados bajo la acción de un campo eléctrico débil y a presión atmosférica. Esta técnica cuenta con una instrumentación de reducidas dimensiones, portátil, sensible, de respuesta rápida y coste relativamente bajo. Sin embargo, la principal limitación de la IMS reside en su incompatibilidad con el análisis directo de muestras líquidas. Por tanto, ha sido necesario el diseño de un inyector donde el IL queda atrapado mientras que los analitos son volatilizados y dirigidos hacia el detector [150–152]. En cuanto a la determinación de especies inorgánicas, diferentes técnicas de LPME basadas en ILs (i.e., IL-SDME e in-situ IL-DLLME) han sido acopladas con técnicas espectrométricas de absorción [153,154] o de emisión [155] atómica basadas en un atomizador de bobina de wolframio. Tras la extracción, la gota enriquecida de IL es diluida


83

con un disolvente orgánico (e.g., tetrahidrofurano, acetonitrilo) para disminuir su viscosidad y, a continuación, es directamente depositada sobre la bobina de wolframio para el análisis [153–155]. Por otro lado, nuestro grupo de investigación propuso el acoplamiento de técnicas como la SDME o la DLLME con la espectrometría de plasmas inducidos por láser (laser-induced

breakdown spectrometry, LIBS), como sistema de detección con opciones de miniaturización y portabilidad [156–159]. Además, este novedoso acoplamiento ha permitido mejorar de forma significativa la sensibilidad y los límites de detección de la técnica LIBS. Por último, conviene mencionar la utilización de los micro-espectrofotómetros de UV, los cuales incluyen cubetas de reducidas dimensiones adecuadas para el análisis de los pocos microlitros de fase extractante utilizados con las técnicas de LPME [160–164].

En la presente memoria se proponen los electrodos serigrafiados (screen-

printed electrodes, SPELs) como sistemas de detección electroquímica en formato miniaturizado para la técnica de DLLME. Estos dispositivos suponen una alternativa interesante al tratarse de una tecnología muy desarrollada, fácilmente disponible y de bajo coste. Como únicos antecedentes podemos señalar los trabajos de investigación presentados por Ahmar et al. en los que diferentes SPELs han sido utilizados para analizar la fase aceptora tras la EME de sustancias de abuso desde muestras de orina y plasma sanguíneo [165–167].

II.2.1. Electrodos serigrafiados

Desde los años noventa, la técnica de serigrafiado utilizada por la ingeniería electrónica para la producción de circuitos impresos se ha ido adaptando a la producción de transductores electroquímicos de bajo coste y reducidas dimensiones conocidos como SPELs [168–170]. Estos dispositivos poseen una configuración compacta que, en el sistema de tres electrodos, incluye un electrodo de trabajo (working electrode, WE), un electrodo de referencia (reference electrode, RE) y un contraelectrodo (counter electrode, CE) (Figura II.41). Dicha configuración permite medir la concentración de un analito confinado en el pequeño volumen de una gota, y su reducido tamaño hace que puedan formar parte de sistemas portátiles. Además, al tratarse de

Electrodos serigrafiados

84

dispositivos de bajo coste, pueden ser desechados tras un único uso evitando así los problemas más habituales de los electrodos sólidos clásicos como son los efectos de memoria y la necesidad de un proceso de limpieza para su reutilización.

Figura II.41. Electrodo serigrafiado comercial de la empresa Dropsens.

Los SPELs son producidos mediante la impresión capa a capa de un fluido tixotrópico (i.e., tinta) sobre un sustrato sólido a través de una red o malla que define la forma y el tamaño final del electrodo [168–170]. El fluido tixotrópico posee una elevada viscosidad la cual disminuye al ser sometido a un esfuerzo de cizalla con una escobilla de goma, permitiendo que penetre a través de la malla y se deposite sobre el sustrato sólido. Una vez en contacto con dicho sustrato, el fluido recupera su viscosidad inicial adquiriendo la forma y el tamaño establecidos por la malla. La Figura II.42 muestra un esquema básico del proceso de serigrafiado de electrodos. Entre la impresión de dos capas consecutivas de tinta, es necesario llevar a cabo un tratamiento térmico para el curado de la misma. Las distintas partes que constituyen un SPEL (i.e., WE, CE, RE, circuito, etc.) son obtenidas utilizando diferentes mallas y/o tintas. Una vez fabricado todo el dispositivo, se aplica una capa protectora final para aislar el circuito conductor.

WE

CE

RE

Conexión del WE

Conexión del RE

Conexión del CE


85

Figura II.42. Representación esquemática del proceso de serigrafiado de electrodos. (Fuente: referencia [168]).

De acuerdo con el procedimiento descrito, la fabricación de SPELs cuenta

con cuatro etapas fundamentales:

(i) Elección de la malla que va a definir la geometría y el tamaño del

electrodo. Los SPELs pueden ser diseñados con diferentes tamaños y formas. Los SPELs comerciales más comunes poseen un WE del orden de los milímetros (Figura II.41). Sin embargo es posible reducir el tamaño de dicho electrodo llegando a dimensiones micrométricas o incluso nanométricas. A estas escalas, los SPELs ofrecen una mejora en los límites de detección y en la relación señal/ruido pero su fabricación es menos reproducible [170]. En cuanto a la forma, las más comunes son las de disco, anillo o banda. (ii) Selección y preparación de la tinta. La formulación exacta de las tintas utilizadas en la fabricación de SPELs comerciales es información protegida por parte de la empresa productora. No obstante, es conocido que dichas tintas están formadas por una gran variedad de sustancias que incluyen el material electródico (e.g., carbono, oro, plata), plastificantes, disolventes y aglutinantes. Todas estas sustancias (tipo y proporción) afectan al comportamiento final del dispositivo. Los SPELs con WE de carbono son los más utilizados debido a sus buenas características electroquímicas (i.e., baja corriente de fondo, amplia ventana de potenciales), estabilidad química y bajo coste [170]. Sin embargo

Malla Tinta

Tinta sobre el sustrato

Pincel

Soporte Sustrato


86

también se han empleado otros materiales como el oro o el platino, aunque su uso está menos extendido debido a su mayor precio. El CE es normalmente fabricado del mismo material que el WE y para el RE o electrodo de pseudoreferencia, se utilizan habitualmente tintas basadas en Ag o mezclas Ag/AgCl. En cuanto a los aditivos que acompañan a estos materiales, los plastificantes (e.g., 2-nitrofenil octil éter) son utilizados para modificar la viscosidad de la tinta; los disolventes (e.g., ciclohexanona) permiten mantenerla en estado líquido antes de la impresión; y un polímero aglutinante (e.g., policloruro de vinilo, acetato de celulosa) es utilizado para facilitar la fijación de la tinta al sustrato sólido [170]. (iii) Serigrafiado. La impresión de la tinta, según el procedimiento descrito, es realizada sobre un sustrato sólido rígido, normalmente de cerámica o plástico. El coste de este último sustrato es inferior y las tintas de carbono se adhieren mejor sobre el mismo. No obstante, en los últimos años se han desarrollado novedosos SPELs sobre sustratos de papel [171] o elásticos [172]. Estos últimos dispositivos son compatibles con superficies como la piel, pudiendo integrarse en la misma y llevar a cabo el monitoreo continuo y en tiempo real de un determinado analito de interés (e.g., glucosa, contaminantes, etc.) [172,173]. (iv) Secado y curado de la tinta. El tratamiento térmico para el secado y curado de la tinta debe hacerse bajo condiciones controladas de tiempo y temperatura, ya que esta etapa puede afectar al comportamiento electroquímico del sensor resultante.

La técnica de serigrafiado ofrece numerosas ventajas entre las que podemos mencionar la flexibilidad de diseño y la posibilidad de automación del proceso. En cuanto a la flexibilidad de diseño, cabe destacar que no solo es posible fabricar dispositivos con electrodos de trabajo de diferente geometría, sino que el diseño completo de la celda electroquímica puede adaptarse a las necesidades concretas del análisis variando el número de electrodos, el volumen, etc. Por otro lado, una vez seleccionadas las mallas y las tintas deseadas, la producción de SPELs puede realizarse en masa a gran escala y con un coste relativamente bajo. Además, otro de los principales atractivos que


87

presentan los SPELs reside en la facilidad de modificar el WE, ya que permite ampliar sus aplicaciones y obtener dispositivos muy específicos en función del analito de interés. Esta gran versatilidad resulta muy interesante y valiosa en el campo de la investigación. La modificación del WE puede realizarse durante el proceso de fabricación, alterando la composición de la tinta, o sobre el dispositivo final mediante la deposición de nanomateriales, enzimas, polímeros, etc., sobre la superficie de dicho electrodo.

II.2.1.1. Técnicas electroanalíticas

Las técnicas electroanalíticas establecen la relación entre las propiedades eléctricas y químicas de las especies que intervienen en una determinada reacción. Estas técnicas ofrecen una respuesta rápida, sensible y reproducible utilizando una instrumentación sencilla, de bajo coste y con opciones de portabilidad, por lo que su utilización resulta muy ventajosa. A continuación, se describirán brevemente las técnicas electroanalíticas utilizadas en la presente memoria para el estudio y la determinación de diferente analitos de interés mediante SPELs.

II.2.1.1.1. Voltamperometría cíclica

La voltamperometría o voltametría cíclica es probablemente la técnica electroanalítica mas eficiente y versátil para el estudio mecanístico de las reacciones electródicas y suele ser el primer experimento que se lleva a cabo en un estudio electroquímico. La importancia de la voltamperometría cíclica reside en la capacidad de proporcionar de manera rápida información sobre la termodinámica del proceso redox, sobre la cinética de la reacción heterogénea de transferencia electrónica y sobre posibles reacciones químicas o procesos de adsorción acoplados. Además, permite la localización rápida de los potenciales a los que tiene lugar la oxidación/reducción de la especie electroactiva [174,175].

La voltamperometría cíclica consiste en la aplicación de un barrido lineal de potencial sobre el WE desde un potencial inicial (Ei) hasta un potencial


88

máximo (Em), al cual se invierte la dirección del barrido hasta alcanzar un potencial final (Ef). De este modo, se aplica un señal de excitación triangular como la que se muestra en la Figura II.43a, donde la pendiente corresponde a la velocidad de barrido utilizada. La señal de excitación puede repetirse una o varias veces (i.e., ciclos) dependiendo de la información deseada.

Figura II.43. (a) Señal de excitación triangular utilizada en voltamperometría cíclica. (b) Voltamperograma típico de una especie redox en disolución.

Para obtener un voltamperograma o voltagrama cíclico, se mide la corriente que pasa por el WE durante el barrido de potencial. La Figura II.43b muestra la respuesta típica de un par redox reversible durante un ciclo de barrido. La forma característica de este tipo de voltamperogramas es debida a la capa de difusión formada en las cercanías del electrodo y puede entenderse haciendo un estudio minucioso de los perfiles de concentración-distancia (Figura II.44) [174,175]. Cada perfil de concentración-distancia señalado con una letra (a-h) se corresponde con el punto señalado con esa misma letra en el voltamperograma de la Figura II.43b. Si se considera un sistema reversible en el que inicialmente sólo se encuentra en disolución la especie reducida (R), durante el barrido directo se alcanza un valor de potencial suficientemente positivo como para provocar la oxidación de dicha especie. Esta oxidación da lugar a una corriente anódica que aumenta rápidamente hasta que la concentración de R en la superficie del electrodo se aproxima a cero. La corriente decae a continuación a medida que se produce el agotamiento de R en las cercanías del electrodo debido a su conversión en la especie oxidada (O).

Ciclo 1

Ei

Em

Ef

Tiempo

Pote

ncia

l

a

b

c

d e

f g

h

R → O + ne-

Potencial

O + ne- → R C

orrie

nte

(a) (b)


89

Este agotamiento de R y acumulación de O cerca del electodo se ilustra por medio de los perfiles de concentración-distancia mostrados en la Figura II.44a. Después de invertir la dirección del barrio de potencial, la oxidación prosigue hasta que el potencial aplicado se hace suficientemente negativo como para causar la reducción de la especie O acumulada. Dicha reducción viene dada por la aparición de una corriente catódica. De nuevo, dicha corriente aumenta a medida que el potencial se hace menos positivo hasta que la reducción de O provoca su agotamiento en la región cercana al electrodo, produciéndose por tanto la aparición de un pico de corriente y su posterior disminución. Los perfiles de concentración-distancia de R y O durante el barrido inverso se muestran en la Figura II.44b.

Figura II.44. (a) Perfil concentración-distancia durante el barrido directo de potencial. (b) Perfil de concentración-distancia al invertir la dirección de barrido. (Fuente: referencias [174,175]).

A parte del proceso faradaico descrito, la voltamperometría cíclica está caracterizada por una corriente de fondo o corriente capacitiva. Dicha corriente es debida a la formación de una doble capa de especies cargadas y/o dipolos orientados en la interfase electródica que compensan el exceso de carga que se

(a)

(b)

Con

cent

raci

ón

Distancia


90

genera en la superficie del electrodo [174,175]. La medida de las intensidades de pico del proceso faradaico requiere por tanto del trazado de una línea base para la sustracción de la corriente de fondo relacionada con los procesos capacitivos. No obstante, conviene señalar que la mayor utilidad de la voltamperométria cíclica radica en su empleo como herramienta para obtener información cualitativa, mientras que otras técnicas como las de impulsos ofrecen una mayor sensibilidad para las medidas cuantitativas.

II.2.1.1.2. Técnicas voltamperométricas de impulsos

Las técnicas voltamperométricas de impulsos tienen como objetivo la disminución de los límites de detección mediante el aumento de la relación entre la corriente faradaica y la no faradaica o capacitiva. Después de la aplicación de un impulso de potencial, la corriente capacitiva disminuye rápidamente (exponencialmente) mientras que la corriente faradaica lo hace más lentamente. De este modo, si se mide la corriente cerca del tiempo final del impulso, dicha medida corresponderá prácticamente en su totalidad con la corriente faradaica [174,175].

La diferencia entre las diversas técnicas voltamperométricas de impulsos reside en la señal de excitación aplicada y en la medida de la corriente.

II.2.1.1.2.1. Voltamperometría diferencial de impulsos

En la voltamperometría diferencial de impulsos (differential pulse

voltammetry, DPV) la señal de excitación consiste en la aplicación de impulsos de magnitud constante sobre una rampa lineal de potencial base (Figura II.45a). La selección de la amplitud del impulso y de la velocidad de barrido requiere una situación de compromiso entre sensibilidad, resolución y velocidad del análisis. Para la obtención de la curva voltamperométrica, la corriente es medida dos veces: una justo antes de la aplicación del impulso (τ1) y otra al final del tiempo de vida del mismo (τ2). La primera corriente es sustraída instrumentalmente a la segunda y esta diferencia [Δi=i(τ2)-i(τ1)] es representada frente al potencial aplicado [174,175]. La Figura II.45b muestra


91

un voltamperograma diferencial de impulsos típico, donde la altura de pico es directamente proporcional a la concentración de la especie redox.

Figura II.45. (a) Señal de excitación y (b) voltamperograma típico de la voltamperometría diferencial de impulsos.

II.2.1.1.2.2. Voltamperometría de onda cuadrada

La señal de excitación aplicada en voltamperometría de onda cuadrada (square-wave voltammetry, SWV) consiste en la combinación de una modulación de onda cuadrada de gran amplitud con una rampa de potencial en forma de escalera (Figura II.46). La amplitud de la modulación de la onda cuadrada (Esw) es tan grade que, si se supone una reacción inicial R O + ne-, los impulsos inversos causan la reducción del producto que se genera en los impulsos directos para volver a dar R, resultando así una corriente inversa anódica. Para la obtención de la curva voltamperométrica, la corriente es medida dos veces: una al final del impulso directo (τ1) y otra al final del impulso inverso (τ2). La diferencia entre ambas corrientes (i.e., corriente neta) se representa frente al potencial base en escalera obteniendo voltamperogramas en forma de pico muy similares al mostrado en la Figura II.45b. La corriente neta es más grande que la corriente directa o que la corriente inversa, ya que se trata de la diferencia algebraica entre ellas, por lo que los voltamperogramas que se obtienen ofrecen una excelente sensibilidad. Además, otra de las mayores ventajas de la SWV es la posibilidad de barrer el intervalo de potenciales de interés a una gran velocidad.

(b) (a) Tiempo entre

impulsos Duración de un impulso

Tiempo

Pote

ncia

l

τ1

τ2

Potencial

Cor

rient

e

Intensidad del impulso


92

Figura II.46. Señal de excitación en voltamperometría de onda cuadrada: Esw, amplitud de la onda cuadrada; y ΔEs, altura de escalón.

II.2.1.1.3. Voltamperometría de redisolución anódica

Las técnicas voltamperométricas de redisolución están basadas en la preconcentración del analito presente en disolución sobre la superficie del electrodo y en su posterior redisolución con una técnica voltamperométrica. La mayor ventaja de estas técnicas radica en la precocentración del analito objeto de estudio en factores que oscilan entre 100 y más de 1000, lo que las hace especialmente adecuadas para el análisis a niveles traza [174,175]. Existen diversas modalidades de técnicas voltamperométricas de redisolución siendo la más común la voltamperometría de redisolución anódica (anodic stripping

voltammetry, ASV).

La ASV se usa generalmente para la determinación de iones metálicos. En ella, la preconcentración se lleva a cabo mediante electrodeposición catódica y, tras un tiempo preseleccionado, los metales electrodepositados son redisueltos

τ1

τ2

Periodo

Esw ΔEs

Tiempo

Pote

ncia

l


93

aplicando un barrido anódico de potencial hacia valores más positivos que el de deposición (Figura II.47). Dicho barrido puede realizarse utilizando técnicas voltamperométricas de barrido lineal o de impulsos (e.g., DPV, SWV). La corriente del pico resultante será proporcional a la concentración de analito y dependerá además de varios factores relacionados con la etapa de deposición (e.g., tiempo de deposición) y con la de redisolución (e.g., velocidad de barrido, intensidad del impulso).

Figura II.47. Esquema del programa de potencial aplicado en la voltamperometría de redisolución anódica.

Tiempo

Pote

ncia

l

Electrodeposición

M+n + ne- M

M M+n + ne-

REFERENCIAS

Referencias

97

[1] M. Valcárcel, A modern definition of analytical chemistry, Trends Anal. Chem. 16 (1997) 124–131.

[2] M. Valcárcel, A. Ríos, The analytical problem, Trends Anal. Chem. 16 (1997) 385–393.

[3] C.E. Domini, D. Hristozov, B. Almagro, S. Prats, A. Canals, Sample preparation for chromatographic analysis of environmental samples, in: L.M. Nollet (Ed.), Chromatografic Analysis of the Environment, 3rd ed., Taylor and Francis Inc., Boca Raton (USA), 2006: pp. 31–131.

[4] R. Cela, R.A. Lorenzo, M. del C. Casais, Técnicas de Separación en Química Analítica, 1a ed., Editorial Síntesis, Madrid (España), 2002.

[5] C.L. Arthur, J. Pawliszyn, Solid phase microextraction with thermal desorption using fused silica optical fibers, Anal. Chem. 62 (1990) 2145–2148.

[6] J. Plotka-Wasylka, N. Szczepanska, M. de la Guardia, J. Namiesnik, Miniaturized solid-phase extraction techniques, Trends Anal. Chem. 73 (2015) 19–38.

[7] R. Eisert, J. Pawliszyn, Automated in-tube solid-phase microextraction coupled to high-performance liquid chromatography, Anal. Chem. 69 (1997) 3140–3147.

[8] M. Fernández-Amado, M.C. Prieto-Blanco, P. López-Mahía, S. Muniategui-Lorenzo, D. Prada-Rodríguez, Strengths and weaknesses of in-tube solid-phase microextraction: A scoping review, Anal. Chim. Acta 906 (2016) 41–57.

[9] H. Lord, J. Pawliszyn, Evolution of solid-phase microextraction technology, J. Chromatogr. A 885 (2000) 153–193.

[10] E. Baltussen, P. Sandra, F. David, C. Cramers, Stir bar sorptive extraction (SBSE), a novel extraction technique for aqueous samples: Theory and principles, J. Microcolumn Sep. 11 (1999) 737–747.

[11] M. Šafaříková, I. Šafařík, Magnetic solid-phase extraction, J. Magn. Magn. Mater. 194 (1999) 108–112.

[12] I. Bruheim, X. Liu, J. Pawliszyn, Thin-film microextraction, Anal. Chem. 75 (2003) 1002–1010.

[13] R. Jiang, J. Pawliszyn, Thin-film microextraction offers another geometry for solid-phase microextraction, Trends Anal. Chem. 39

Referencias

98

(2012) 245–253. [14] M. Abdel-Rehim, New trend in sample preparation: on-line

microextraction in packed syringe for liquid and gas chromatography applications I. Determination of local anaesthetics in human plasma samples using gas chromatography-mass spectrometry, J. Chromatogr. B 801 (2004) 317–321.

[15] C. Basheer, A.A. Alnedhary, B.S.M. Rao, S. Valliyaveettil, H.K. Lee, Development and application of porous membrane-protected carbon nanotube micro-solid-phase extraction combined with gas chromatography/mass spectrometry, Anal. Chem. 78 (2006) 2853–2858.

[16] W.H. Tsai, T.-C. Huang, J.-J. Huang, Y.-H. Hsue, H.-Y. Chuang, Dispersive solid-phase microextraction method for sample extraction in the analysis of four tetracyclines in water and milk samples by high-performance liquid chromatography with diode-array detection, J. Chromatogr. A 1216 (2009) 2263–2269.

[17] M. Anastassiades, S.J. Lehotay, D. Stajnbaher, F.J. Schenck, Fast and easy multiresidue method employing acetonitrile extraction/partitioning and “dispersive solid-phase extraction” for determination of pesticides residues in produce, J. AOAC Int. 86 (2003) 412–431.

[18] M. Cruz-Vera, R. Lucena, S. Cárdenas, M. Valcárcel, Sample treatments based on dispersive (micro)extraction, Anal. Methods 3 (2011) 1719–1728.

[19] T. Khezeli, A. Daneshfar, Development of dispersive micro-solid phase extraction based on micro and nano sorbents, Trends Anal. Chem. 89 (2017) 99–118.

[20] M. Khan, E. Yilmaz, M. Soylak, Vortex assisted magnetic solid phase extraction of lead(II) and cobalt(II) on silica coated magnetic multiwalled carbon nanotubes impregnated with 1-(2-pyridylazo)-2-naphthol, J. Mol. Liq. 224 (2016) 639–647.

[21] R.M. Peña-Crecente, C. Gutierrez-Lovera, J. Barciela-García, C. Herrero-Latorre, S. García-Martín, Ultrasound-assisted magnetic solid-phase extraction for the determination of some transition metals in Orujo spirit samples by capillary electrophoresis, Food Chem. 190 (2016) 263–269.

Referencias

99

[22] Á. Ríos, M. Zougagh, Recent advances in magnetic nanomaterials for improving analytical processes, Trends Anal. Chem. 84 (2016) 72–83.

[23] L. Xie, R. Jiang, F. Zhu, H. Liu, G. Ouyang, Application of functionalized magnetic nanoparticles in sample preparation, Anal. Bioanal. Chem. 406 (2014) 377–399.

[24] C. Herrero-Latorre, J. Barciela-García, S. García-Martín, R.M. Peña-Crecente, J. Otárola-Jiménez, Magnetic solid-phase extraction using carbon nanotubes as sorbents: A review, Anal. Chim. Acta 892 (2015) 10–26.

[25] A. Mehdinia, H. Khani, S. Mozaffari, Fibers coated with a graphene-polyaniline nanocomposite for the headspace solid-phase microextraction of organochlorine pesticides from seawater samples, Microchim. Acta 181 (2014) 89–95.

[26] Y. Moliner-Martinez, R. Herráez-Hernández, J. Verdú-Andrés, C. Molins-Legua, P. Campíns-Falcó, Recent advances of in-tube solid-phase microextraction, Trends Anal. Chem. 71 (2015) 205–213.

[27] M.M. Moein, A. Abdel-Rehim, M. Abdel-Rehim, Microextraction by packed sorbent (MEPS), Trends Anal. Chem. 67 (2015) 34–44.

[28] Q. Wu, G. Zhao, C. Feng, C. Wang, Z. Wang, Preparation of a graphene-based magnetic nanocomposite for the extraction of carbamate pesticides from environmental water samples, J. Chromatogr. A 1218 (2011) 7936–7942.

[29] L. Costa dos Reis, L. Vidal, A. Canals, Graphene oxide/Fe3O4 as sorbent for magnetic solid-phase extraction coupled with liquid chromatography to determine 2,4,6-trinitrotoluene in water samples, Anal. Bioanal. Chem. 409 (2017) 2665–2674.

[30] A. Speltini, M. Sturini, F. Maraschi, A. Profumo, Recent trends in the application of the newest carbonaceous materials for magnetic solid-phase extraction of environmental pollutants, Trends Environ. Anal. Chem. 10 (2016) 11–23.

[31] F. Maya, C.P. Cabello, R.M. Frizzarin, J.M. Estela, G. Turnes, V. Cerdà, Magnetic solid-phase extraction using metal-organic frameworks (MOFs) and their derived carbons, Trends Anal. Chem. 90 (2017) 142–152.

Referencias

100

[32] P. Rocío-Bautista, I. Pacheco-Fernández, J. Pasán, V. Pino, Are metal-organic frameworks able to provide a new generation of solid-phase microextraction coatings? - A review, Anal. Chim. Acta 939 (2016) 26–41.

[33] S.-H. Huo, X.-P. Yan, Facile magnetization of metal-organic framework MIL-101 for magnetic solid-phase extraction of polycyclic aromatic hydrocarbons in environmental water samples, Analyst 137 (2012) 3445–3451.

[34] M.G. Valdés, A.I. Pérez-Cordoves, M.E. Díaz-García, Zeolites and zeolite-based materials in analytical chemistry, Trends Anal. Chem. 25 (2006) 24–30.

[35] Structure Commision of the International Zeolite Association (January, 2017), http://www.iza-structure.org/.

[36] E.M. Flanigen, R.W. Broach, S.T. Wilson, Introduction, in: S. Kulprathipanja (Ed.), Zeolites in Industrial Separation and Catalysis, 1st ed., Wiley-VCH, Des Plaines (USA), 2010: pp. 1–26.

[37] M. Zaarour, B. Dong, I. Naydenova, R. Retoux, S. Mintova, Progress in zeolite synthesis promotes advanced applications, Micropor. Mesopor. Mater. 189 (2014) 11–21.

[38] A. Chalabiani, A.A. Matin, K. Farhadi, Zeolite-SiC in PVC matrix as a new SPME fiber for gas chromatographic determination of BTEX in water and soil samples, J. Chinese Chem. Soc. 59 (2012) 1080–1085.

[39] T.P. Lee, B. Saad, E.P. Ng, B. Salleh, Zeolite Linde Type L as micro-solid phase extraction sorbent for the high performance liquid chromatography determination of ochratoxin A in coffee and cereal, J. Chromatogr. A 1237 (2012) 46–54.

[40] W.B. Wilson, A.A. Costa, H. Wang, J.A. Dias, S.C.L. Dias, A.D. Campiglia, Analytical evaluation of BEA zeolite for the pre-concentration of polycyclic aromatic hydrocarbons and their subsequent chromatographic analysis in water samples, Anal. Chim. Acta 733 (2012) 103–109.

[41] T. Pasinli, E. Henden, Solid phase extraction of zearalenone (ZEN) from beer samples using natural zeolite clinoptilolite and organo-zeolites prior to HPLC determination, Ekoloji 22 (2013) 57–66.

Referencias

101

[42] S. Goda, R. Selyanchyn, T. Nozoe, H. Matsui, S.W. Lee, Development of a thin-film microextraction device based on ZSM-5/Tenax TA for VOC detection in liquid samples, J. Anal. Bioanal. Tech. (2014) 004/1-004/7.

[43] H. Wang, Z. Li, Q. Niu, J. Ma, Q. Jia, Preparation of a zeolite-modified polymer monolith for identification of synthetic colorants in lipsticks, Appl. Surf. Sci. 353 (2015) 1326–1333.

[44] M.C. Dumlao, L.E. Jeffress, J.J. Gooding, W.A. Donald, Solid-phase microextraction low temperature plasma mass spectrometry for the direct and rapid analysis of chemical warfare simulants in complex mixtures, Analyst 141 (2016) 3714–3721.

[45] P. Salisaeng, P. Arnnok, N. Patdhanagul, R. Burakham, Vortex-assisted dispersive micro-solid phase extraction using CTAB-modified zeolite NaY sorbent coupled with HPLC for the determination of carbamate insecticides, J. Agric. Food Chem. 64 (2016) 2145–2152.

[46] A. Mollahosseini, M. Toghroli, M. Kamankesh, Zeolite/Fe3O4 as a new sorbent in magnetic solid-phase extraction followed by gas chromatography for determining phthalates in aqueous samples, J. Sep. Sci. 38 (2015) 3750–3757.

[47] S. Zhou, Z. Li, X. Lv, B. Hu, Q. Jia, Preconcentration of synthetic phenolic antioxidants by using magnetic zeolites derived with carboxylatocalix[4]arenes combined with high performance liquid chromatography, Analyst 140 (2015) 5944–5952.

[48] S. Ansari, M. Karimi, Recent configurations and progressive uses of magnetic molecularly imprinted polymers for drug analysis, Talanta 167 (2017) 470–485.

[49] D. Huang, C. Deng, X. Zhang, Functionalized magnetic nanomaterials as solid-phase extraction adsorbents for organic pollutants in environmental analysis, Anal. Methods 6 (2014) 7130–7141.

[50] I.S. Ibarra, J.A. Rodriguez, C.A. Galán-Vidal, A. Cepeda, J.M. Miranda, Magnetic solid phase extraction applied to food analysis, J. Chem. 2015 (2015) 1–13.

[51] V.V. Tolmacheva, V.V. Apyari, E.V. Kochuk, S.G. Dmitrienko, Magnetic adsorbents based on iron oxide nanoparticles for the extraction

Referencias

102

and preconcentration of organic compounds, J. Anal. Chem. 71 (2016) 321–338.

[52] I. Vasconcelos, C. Fernandes, Magnetic solid phase extraction for determination of drugs in biological matrices, Trends Anal. Chem. 89 (2017) 41–52.

[53] M. Wierucka, M. Biziuk, Application of magnetic nanoparticles for magnetic solid-phase extraction in preparing biological, environmental and food samples, Trends Anal. Chem. 59 (2014) 50–58.

[54] J.M. Kokosa, Advances in solvent-microextraction techniques, Trends Anal. Chem. 43 (2013) 2–13.

[55] S. Liu, P.K. Dasgupta, Liquid Droplet. A renewable gas samplig interface, Anal. Chem. 67 (1995) 2042–2049.

[56] A.A. Cardoso, P.K. Dasgupta, Analytical chemistry in a liquid film/droplet, Anal. Chem. 67 (1995) 2562–2566.

[57] H. Liu, P.K. Dasgupta, Analytical chemistry in a drop. Solvent extraction in a microdrop, Anal. Chem. 68 (1996) 1817–1821.

[58] M.A. Jeannot, F.F. Cantwell, Solvent microextraction into a single drop, Anal. Chem. 68 (1996) 2236–2240.

[59] M.A. Jeannot, F.F. Cantwell, Mass transfer characteristics of solvent extraction into a single drop at the tip of a syringe needle, Anal. Chem. 69 (1997) 235–239.

[60] M. Ma, F.F. Cantwell, Solvent microextraction with simultaneous back-extraction for sample cleanup and preconcentration: Quantitative extraction, Anal. Chem. 70 (1998) 3912–3919.

[61] Z. Fan, X. Liu, Determination of methylmercury and phenylmercury in water samples by liquid-liquid-liquid microextraction coupled with capillary electrophoresis, J. Chromatogr. A 1180 (2008) 187–192.

[62] W. Gao, G. Chen, T. Chen, X. Zhang, Y. Chen, Z. Hu, Directly suspended droplet microextraction combined with single drop back-extraction as a new approach for sample preparation compatible with capillary electrophoresis, Talanta 83 (2011) 1673–1679.

[63] S. Pedersen-Bjergaard, K.E. Rasmussen, Liquid-liquid-liquid microextraction for sample preparation of biological fluids prior to capillary electrophoresis, Anal. Chem. 71 (1999) 2650–2656.

Referencias

103

[64] T.S. Ho, T.G. Halvorsen, S. Pedersen-Bjergaard, K.E. Rasmussen, Liquid-phase microextraction of hydrophilic drugs by carrier-mediated transport, J. Chromatogr. A 998 (2003) 61–72.

[65] W. Liu, H.K. Lee, Continous flow microextraction exceeding 1000-fold concentration of dilute analytes, Anal. Chem. 72 (2000) 4462–4467.

[66] Y. Liu, Y. Hashi, J.-M. Lin, Continuous-flow microextraction and gas chromatographic-mass spectrometric determination of polycyclic aromatic hydrocarbon compounds in water, Anal. Chim. Acta 585 (2007) 294–299.

[67] L. Xia, B. Hu, Z. Jiang, Y. Wu, Y. Liang, Single-drop microextraction combined with low-temperature electrothermal vaporization ICPMS for the determination of trace Be, Co, Pd, and Cd in biological samples, Anal. Chem. 76 (2004) 2910–2915.

[68] A.L. Theis, A.J. Waldack, S.M. Hansen, M.A. Jeannot, Headspace solvent microextraction, Anal. Chem. 73 (2001) 5651–5654.

[69] A. Tankeviciute, R. Kazlauskas, V. Vickackaite, Headspace extraction of alcohols into a single drop, Analyst 126 (2001) 1674–1677.

[70] X. Jiang, H.K. Lee, Solvent bar microextraction, Anal. Chem. 76 (2004) 5591–5596.

[71] S. Pedersen-Bjergaard, K.E. Rasmussen, Electrokinetic migration across artificial liquid membranes: New concept for rapid sample preparation of biological fluids, J. Chromatogr. A 1109 (2006) 183–190.

[72] M. Rezaee, Y. Assadi, M.-R. Milani Hosseini, E. Aghaee, F. Ahmadi, S. Berijani, Determination of organic compounds in water using dispersive liquid-liquid microextraction, J. Chromatogr. A 1116 (2006) 1–9.

[73] L. Yangcheng, L. Quan, L. Guangsheng, D. Youyuan, Directly suspended droplet microextraction, Anal. Chim. Acta 566 (2006) 259–264.

[74] G. Mingyuan, L. Yangcheng, L. Guangsheng, Directly suspended droplet microextraction in a rotating vial, Anal. Chim. Acta 648 (2009) 123–127.

[75] M.R. Khalili Zanjani, Y. Yamini, S. Shariati, J.A. Jönsson, A new liquid-phase microextraction method based on solidification of floating organic drop, Anal. Chim. Acta 585 (2007) 286–293.

Referencias

104

[76] A. Spietelun, L. Marcinkowski, M. de la Guardia, J. Namieśnik, Green aspects, developments and perspectives of liquid phase microextraction techniques, Talanta 119 (2014) 34–45.

[77] J. Liu, G. Jiang, Y. Chi, Y. Cai, Q. Zhou, J.-T. Hu, Use of ionic liquids for liquid-phase microextraction of polycyclic aromatic hydrocarbons, Anal. Chem. 75 (2003) 5870–5876.

[78] P. Sun, D.W. Armstrong, Ionic liquids in analytical chemistry, Anal. Chim. Acta 661 (2010) 1–16.

[79] T.D. Ho, C. Zhang, L.W. Hantao, J.L. Anderson, Ionic liquids in analytica chemistry: Fundamentals, advances, and perspectives, Anal. Chem. 86 (2014) 262–286.

[80] E. Aguilera-Herrador, R. Lucena, S. Cárdenas, M. Valcárcel, The roles of ionic liquids in sorptive microextraction techniques, Trends Anal. Chem. 29 (2010) 602–616.

[81] Q. Zhou, H. Bai, G. Xie, J. Xiao, Temperature-controlled ionic liquid dispersive liquid phase micro-extraction, J. Chromatogr. A 1177 (2008) 43–49.

[82] M. Baghdadi, F. Shemirani, In situ solvent formation microextraction based on ionic liquids: A novel sample preparation technique for determination of inorganic species in saline solutions, Anal. Chim. Acta 634 (2009) 186–191.

[83] C. Yao, J.L. Anderson, Dispersive liquid-liquid microextraction using an in situ metathesis reaction to form an ionic liquid extraction phase for the preconcentration of aromatic compounds from water, Anal. Bioanal. Chem. 395 (2009) 1491–1502.

[84] T.P.T. Pham, C.-W. Cho, Y.-S. Yun, Environmental fate and toxicity of ionic liquids: A review, Water Res. 44 (2010) 352–372.

[85] D. Ott, D. Kralisch, A. Stark, Perspectives of ionic liquids as environmentally benign substitutes for molecular solvents, in: P.T. Anastas (Ed.), Handbook of Green Chemistry, Vol 6: Ionic Liquids, 1st ed., Wiley-VCH, New York (USA), 2010: pp. 315–340.

[86] K.D. Clark, O. Nacham, J.A. Purslow, S.A. Pierson, J.L. Anderson, Magnetic ionic liquids in analytical chemistry: A review, Anal. Chim. Acta 934 (2016) 9–21.

Referencias

105

[87] Y. Wang, Y. Sun, B. Xu, X. Li, R. Jin, H. Zhang, D. Song, Magnetic ionic liquid-based dispersive liquid-liquid microextraction for the determination of triazine herbicides in vegetable oils by liquid chromatography, J. Chromatogr. A 1373 (2014) 9–16.

[88] Y. Wang, Y. Sun, B. Xu, X. Li, X. Wang, H. Zhang, D. Song, Matrix solid-phase dispersion coupled with magnetic ionic liquid dispersive liquid-liquid microextraction for the determination of triazine herbicides in oilseeds, Anal. Chim. Acta 888 (2015) 67–74.

[89] K.D. Clark, O. Nacham, H. Yu, T. Li, M.M. Yamsek, D.R. Ronning, J.L. Anderson, Extraction of DNA by magnetic ionic liquids: Tunable solvents for rapid and selective DNA analysis, Anal. Chem. 87 (2015) 1552–1559.

[90] M.J. Trujillo-Rodríguez, O. Nacham, K.D. Clark, V. Pino, J.L. Anderson, J.H. Ayala, A.M. Afonso, Magnetic ionic liquids as non-conventional extraction solvents for the determination of polycyclic aromatic hydrocarbons, Anal. Chim. Acta 934 (2016) 106–113.

[91] H. Yu, J. Merib, J.L. Anderson, Faster dispersive liquid-liquid microextraction methods using magnetic ionic liquids as solvents, J. Chromatogr. A 1463 (2016) 11–19.

[92] J. An, K.L. Rahn, J.L. Anderson, Headspace single drop microextraction versus dispersive liquid-liquid microextraction using magnetic ionic liquid extraction solvents, Talanta 167 (2017) 268–278.

[93] X. Wang, G. Xu, P. Chen, X. Liu, Y. Fang, S. Yang, G. Wang, Arsenic speciation analysis in environmental water, sediment and soil samples by magnetic ionic liquid-based air-assisted liquid-liquid microextraction, RSC Adv. 6 (2016) 110247–110254.

[94] G. Shen, H.K. Lee, Headspace liquid-phase microextraction of chlorobenzenes in soil with gas chromatography-electron capture detection, Anal. Chem. 75 (2003) 98–103.

[95] M. Saraji, Dynamic headspace liquid-phase microextraction of alcohols, J. Chromatogr. A 1062 (2005) 15–21.

[96] G. Ouyang, W. Zhao, J. Pawliszyn, Kinetic calibration for automated headspace liquid-phase microextraction, Anal. Chem. 77 (2005) 8122–8128.

Referencias

106

[97] G. Ouyang, W. Zhao, J. Pawliszyn, Automation and optimization of liquid-phase microextraction by gas chromatography, J. Chromatogr. A 1138 (2007) 47–54.

[98] C.R. Schnobrich, M.A. Jeannot, Steady-state kinetic model for headspace solvent microextraction, J. Chromatogr. A 1215 (2008) 30–36.

[99] L. Vidal, C.E. Domini, N. Grané, E. Psillakis, A. Canals, Microwave-assisted headspace single-drop microextration of chlorobenzenes from water samples, Anal. Chim. Acta 592 (2007) 9–15.

[100] H. Xu, Y. Liao, J. Yao, Development of a novel ultrasound-assisted headspace liquid-phase microextraction and its application to the analysis of chlorophenols in real aqueous samples, J. Chromatogr. A 1167 (2007) 1–8.

[101] A. Gjelstad, K.E. Rasmussen, S. Pedersen-Bjergaard, Simulation of flux during electro-membrane extraction based on the Nernst-Planck equation, J. Chromatogr. A 1174 (2007) 104–111.

[102] K.F. Seip, H. Jensen, M.H. Sønsteby, A. Gjelstad, S. Pedersen-Bjergaard, Electromembrane extraction: distribution or electrophoresis?, Electrophoresis 34 (2013) 792–799.

[103] C. Huang, H. Jensen, K.F. Seip, A. Gjelstad, S. Pedersen-Bjergaard, Mass transfer in electromembrane extraction-The link between theory and experiments, J. Sep. Sci. 39 (2016) 188–197.

[104] P. Kubán, P. Boček, The effects of electrolysis on operational solutions in electromembrane extraction: The role of acceptor solution, J. Chromatogr. A 1398 (2015) 11–19.

[105] A. Slampová, P. Kubán, P. Boček, Additional considerations on electrolysis in electromembrane extraction, J. Chromatogr. A 1429 (2016) 364–368.

[106] A. Gjelstad, S. Pedersen-Bjergaard, Recent developments in electromembrane extraction, Anal. Methods 5 (2013) 4549–4557.

[107] M. Rezazadeh, Y. Yamini, S. Seidi, A. Esrafili, Pulsed electromembrane extraction: A new concept of electrically enhanced extraction, J. Chromatogr. A 1262 (2012) 214–218.

[108] A. Slampová, P. Kubáň, P. Boček, Electromembrane extraction using

Referencias

107

stabilized constant d.c. electric current-A simple tool for improvement of extraction performance, J. Chromatogr. A 1234 (2012) 32–37.

[109] S.S.H. Davarani, A. Morteza-Najarian, S. Nojavan, A. Pourahadi, M.B. Abbassi, Two-phase electromembrane extraction followed by gas chromatography-mass spectrometry analysis, J. Sep. Sci. 36 (2013) 736–743.

[110] C. Basheer, J. Lee, S. Pedersen-Bjergaard, K.E. Rasmussen, H.K. Lee, Simultaneous extraction of acidic and basic drugs at neutral sample pH: A novel electro-mediated microextraction approach, J. Chromatogr. A 1217 (2010) 6661–6667.

[111] S. Seidi, Y. Yamini, M. Rezazadeh, A. Esrafili, Low-voltage electrically-enhanced microextraction as a novel technique for simultaneous extraction of acidic and basic drugs from biological fluids, J. Chromatogr. A 1243 (2012) 6–13.

[112] N.J. Petersen, H. Jensen, S.H. Hansen, K.E. Rasmussen, S. Pedersen-Bjergaard, Drop-to-drop microextraction across a supported liquid membrane by an electrical field under stagnant conditions, J. Chromatogr. A 1216 (2009) 1496–1502.

[113] N.J. Petersen, H. Jensen, S.H. Hansen, S.T. Foss, D. Snakenborg, S. Pedersen-Bjergaard, On-chip electro membrane extraction, Microfluid. Nanofluidics 9 (2010) 881–888.

[114] M.D. Ramos-Payán, B. Li, N.J. Petersen, H. Jensen, S.H. Hansen, S. Pedersen-Bjergaard, Nano-electromembrane extraction, Anal. Chim. Acta 785 (2013) 60–66.

[115] H. Xu, Z. Ding, L. Lv, D. Song, Y.-Q. Feng, A novel dispersive liquid-liquid microextraction based on solidification of floating organic droplet method for determination of polycyclic aromatic hydrocarbons in aqueous samples, Anal. Chim. Acta 636 (2009) 28–33.

[116] M.-I. Leong, S.-D. Huang, Dispersive liquid-liquid microextraction method based on solidification of floating organic drop for extraction of organochlorine pesticides in water samples, J. Chromatogr. A 1216 (2009) 7645–7650.

[117] L. Kocúrová, I.S. Balogh, J. Šandrejová, V. Andruch, Recent advances in dispersive liquid-liquid microextraction using organic solvents lighter

Referencias

108

than water. A review, Microchem. J. 102 (2012) 11–17. [118] A. Ballesteros-Gómez, M.D. Sicilia, S. Rubio, Supramolecular solvents

in the extraction of organic compounds. A review, Anal. Chim. Acta 677 (2010) 108–130.

[119] M. Moradi, Y. Yamini, Surfactant roles in modern sample preparation techniques: A review, J. Sep. Sci. 35 (2012) 2319–2340.

[120] M. Saraji, A.A.H. Bidgoli, Dispersive liquid-liquid microextraction using a surfactant as disperser agent, Anal. Bioanal. Chem. 397 (2010) 3107–3115.

[121] V. Andruch, M. Burdel, L. Kocúrová, J. Šandrejová, I.S. Balogh, Application of ultrasonic irradiation and vortex agitation in solvent microextraction, Trends Anal. Chem. 49 (2013) 1–19.

[122] M. Cruz-Vera, R. Lucena, S. Cárdenas, M. Valcárcel, One-step in-syringe ionic liquid-based dispersive liquid-liquid microextraction, J. Chromatogr. A 1216 (2009) 6459–6465.

[123] V. Andruch, C.C. Acebal, J. Škrlíková, H. Sklenářová, P. Solich, I.S. Balogh, F. Billes, L. Kocúrová, Automated on-line dispersive liquid-liquid microextraction based on a sequential injection system, Microchem. J. 100 (2012) 77–82.

[124] M.A. Farajzadeh, D. Djozan, P. Khorram, Development of a new dispersive liquid-liquid microextraction method in a narrow-bore tube for preconcentration of triazole pesticides from aqueous samples, Anal. Chim. Acta 713 (2012) 70–78.

[125] C.K. Zacharis, P.D. Tzanavaras, K. Roubos, K. Dhima, Solvent-based de-emulsification dispersive liquid-liquid microextraction combined with gas chromatography-mass spectrometry for determination of trace organochlorine pesticides in environmental water samples, J. Chromatogr. A 1217 (2010) 5896–5900.

[126] H. Chen, R. Chen, S. Li, Low-density extraction solvent-based solvent terminated dispersive liquid-liquid microextraction combined with gas chromatography-tandem mass spectrometry for the determination of carbamate pesticides in water samples, J. Chromatogr. A 1217 (2010) 1244–1248.

[127] A.N. Anthemidis, K.-I.G. Ioannou, Recent developments in

Referencias

109

homogeneous and dispersive liquid-liquid extraction for inorganic elements determination. A review, Talanta 80 (2009) 413–421.

[128] A. V. Herrera-Herrera, M. Asensio-Ramos, J. Hernández-Borges, M.Á. Rodríguez-Delgado, Dispersive liquid-liquid microextraction for determination of organic analytes, Trends Anal. Chem. 29 (2010) 728–751.

[129] M. Rezaee, Y. Yamini, M. Faraji, Evolution of dispersive liquid-liquid microextraction method, J. Chromatogr. A 1217 (2010) 2342–2357.

[130] A. Zgoła-Grześkowiak, T. Grześkowiak, Dispersive liquid-liquid microextraction, Trends Anal. Chem. 30 (2011) 1382–1399.

[131] H. Yan, H. Wang, Recent development and applications of dispersive liquid-liquid microextraction, J. Chromatogr. A 1295 (2013) 1–15.

[132] M. Saraji, M.K. Boroujeni, Recent developments in dispersive liquid-liquid microextraction, Anal. Bioanal. Chem. 406 (2014) 2027–2066.

[133] P. Viñas, N. Campillo, I. López-García, M. Hernández-Córdoba, Dispersive liquid-liquid microextraction in food analysis. A critical review, Anal. Bioanal. Chem. 406 (2014) 2067–2099.

[134] M.J. Trujillo-Rodríguez, P. Rocío-Bautista, V. Pino, A.M. Afonso, Ionic liquids in dispersive liquid-liquid microextraction, Trends Anal. Chem. 51 (2013) 87–106.

[135] P. Hashemi, F. Raeisi, A.R. Ghiasvand, A. Rahimi, Reversed-phase dispersive liquid-liquid microextraction with central composite design optimization for preconcentration and HPLC determination of oleuropein, Talanta 80 (2010) 1926–1931.

[136] P. Hashemi, F. Nazari Serenjeh, A.R. Ghiasvand, Reversed-phase dispersive liquid-liquid microextraction with multivariate optimization for sensitive HPLC determination of tyrosol and hydroxytyrosol in olive oil, Anal. Sci. 27 (2011) 943–947.

[137] M.P. Godoy-Caballero, M.I. Acedo-Valenzuela, T. Galeano-Díaz, New reversed phase dispersive liquid-liquid microextraction method for the determination of phenolic compounds in virgin olive oil by rapid resolution liquid chromathography with ultraviolet-visible and mass spectrometry detection, J. Chromatogr. A 1313 (2013) 291–301.

[138] N. Goudarzi, S. Farsimadan, M.A. Chamjangali, G.A. Bagherian,

Referencias

110

Development of coupled ultrasound-assisted and reversed-phase dispersive liquid-liquid microextraction before high-performance liquid chromatography for the sensitive determination of Vitamin A and Vitamin E in oil samples, J. Sep. Sci. 38 (2015) 3254–3261.

[139] W.-X. Wang, T.-J. Yang, Z.-G. Li, T.-T. Jong, M.-R. Lee, A novel method of ultrasound-assisted dispersive liquid-liquid microextraction coupled to liquid chromatography-mass spectrometry for the determination of trace organoarsenic compounds in edible oil, Anal. Chim. Acta 690 (2011) 221–227.

[140] M.D. Luque de Castro, F. Priego Capote, Analytical Applications of Ultrasounds, 1st ed., Elsevier B.V., Amsterdam (The Nederlands), 2007.

[141] J. Regueiro, M. Llompart, C. Garcia-Jares, J.C. Garcia-Monteagudo, R. Cela, Ultrasound-assisted emulsification-microextraction of emergent contaminants and pesticides in environmental waters, J. Chromatogr. A 1190 (2008) 27–38.

[142] Y. Picó, Ultrasound-assisted extraction for food and environmental samples, Trends Anal. Chem. 43 (2013) 84–99.

[143] C. Cortada, L. Vidal, A. Canals, Determination of nitroaromatic explosives in water samples by direct ultrasound-assisted dispersive liquid-liquid microextraction followed by gas chromatography-mass spectrometry, Talanta 85 (2011) 2546–2552.

[144] E. Yiantzi, E. Psillakis, K. Tyrovola, N. Kalogerakis, Vortex-assisted liquid-liquid microextraction of octylphenol, nonylphenol and bisphenol-A, Talanta 80 (2010) 2057–2062.

[145] D.C. Montgomery, Design and Analysis of Experiments, 7th ed., Wiley-VCH, New Jersey (USA), 2009.

[146] Y.V. Heyden, C. Hartmann, D.L. Massart, L. Michel, P. Kiechle, F. Erni, H. Fabre, N. Mesplet, Ruggedness tests for a high-performance liquid chromatographic assay: Comparison of an evaluation at two and three levels by using two-level Plackett-Burman designs, Anal. Chim. Acta 897 (1995) 329–338.

[147] S. Sharma, L.T. Tolley, H.D. Tolley, A. Plistil, S.D. Stearns, M.L. Lee, Hand-portable liquid chromatographic instrumentation, J. Chromatogr. A 1421 (2015) 38–47.

Referencias

111

[148] M. Mittermüller, D.A. Volmer, Micro- and nanostructures and their application in gas chromatography, Analyst 137 (2012) 3195–3201.

[149] Y. Peng, D.E. Austin, New approaches to miniaturizing ion trap mass analyzers, Trends Anal. Chem. 30 (2011) 1560–1567.

[150] E. Aguilera-Herrador, R. Lucena, S. Cárdenas, M. Valcárcel, Ionic liquid-based single drop microextraction and room-temperature gas chromatography for on-site ion mobility spectrometric analysis, J. Chromatogr. A 1216 (2009) 5580–5587.

[151] I. Márquez-Sillero, E. Aguilera-Herrador, S. Cárdenas, M. Valcárcel, Determination of 2,4,6-tricholoroanisole in water and wine samples by ionic liquid-based single-drop microextraction and ion mobility spectrometry, Anal. Chim. Acta 702 (2011) 199–204.

[152] I. Márquez-Sillero, S. Cárdenas, M. Valcárcel, Direct determination of 2,4,6-tricholoroanisole in wines by single-drop ionic liquid microextraction coupled with multicapillary column separation and ion mobility spectrometry detection, J. Chromatogr. A 1218 (2011) 7574–7580.

[153] X. Wen, Q. Deng, J. Wang, S. Yang, X. Zhao, A new coupling of ionic liquid based-single drop microextraction with tungsten coil electrothermal atomic absorption spectrometry, Spectrochim. Acta A 105 (2013) 320–325.

[154] X. Wen, Q. Deng, J. Wang, X. Zhao, Application of portable spectrometer for ultra trace metal analysis after ionic liquid based microextration, Anal. Methods 5 (2013) 2978–2983.

[155] L. Vidal, S.G. Silva, A. Canals, J.A. Nóbrega, Tungsten coil atomic emission spectrometry combined with dispersive liquid-liquid microextraction: A synergistic association for chromium determination in water samples, Talanta 148 (2016) 602–608.

[156] I. Gaubeur, M.A. Aguirre, N. Kovachev, M. Hidalgo, A. Canals, Dispersive liquid-liquid microextraction combined with laser-induced breakdown spectrometry and inductively coupled plasma optical emission spectrometry to elemental analysis, Microchem. J. 121 (2015) 219–226.

[157] M.A. Aguirre, E.J. Selva, M. Hidalgo, A. Canals, Dispersive liquid-

Referencias

112

liquid microextraction for metals enrichment: A useful strategy for improving sensitivity of laser-induced breakdown spectroscopy in liquid samples analysis, Talanta 131 (2015) 348–353.

[158] I. Gaubeur, M.Á. Aguirre, N. Kovachev, M. Hidalgo, A. Canals, Speciation of chromium by dispersive liquid-liquid microextraction followed by laser-induced breakdown spectrometry detection (DLLME-LIBS), J. Anal. At. Spectrom. 30 (2015) 2541–2547.

[159] M.A. Aguirre, H. Nikolova, M. Hidalgo, A. Canals, Hyphenation of single-drop microextraction with laser-induced breakdown spectrometry for trace analysis in liquid samples: a viability study, Anal. Methods 7 (2015) 877–883.

[160] A.K.K.V. Pillai, A. Jain, K.K. Verma, Headspace single-drop microextraction and fibre optics-based cuvetteless micro-spectrophotometry for the determination of chloride involving oxidation with permanganate, Talanta 80 (2010) 1816–1822.

[161] F. Pena-Pereira, S. Senra-Ferreiro, I. Lavilla, C. Bendicho, Determination of iodate in waters by cuvetteless UV-vis micro-spectrophotometry after liquid-phase microextraction, Talanta 81 (2010) 625–629.

[162] A. Jain, A.K.K.V. Pillai, N. Sharma, K.K. Verma, Headspace single-drop microextraction and cuvetteless microspectrophotometry for the selective determination of free and total cyanide involving reaction with ninhydrin, Talanta 82 (2010) 758–765.

[163] N. Cabaleiro, F. Pena-Pereira, I. de la Calle, C. Bendicho, I. Lavilla, Determination of triclosan by cuvetteless UV-vis micro-spectrophotometry following simultaneous ultrasound assisted emulsification-microextraction with derivatization: Use of a micellar-ionic liquid as extractant, Microchem. J. 99 (2011) 246–251.

[164] S. Senra-Ferreiro, F. Pena-Pereira, I. Costas-Mora, V. Romero, I. Lavilla, C. Bendicho, Ion pair-based liquid-phase microextraction combined with cuvetteless UV-vis micro-spectrophotometry as a miniaturized assay for monitoring ammonia in waters, Talanta 85 (2011) 1448–1452.

[165] H. Ahmar, A.R. Fakhari, H. Tabani, A. Shahsavani, Optimization of

Referencias

113

electromembrane extraction combined with differential pulse voltammetry using modified screen-printed electrode for the determination of sufentanil, Electrochim. Acta 96 (2013) 117–123.

[166] H. Ahmar, H. Tabani, M.H. Koruni, S.S.H. Davarani, A.R. Fakhari, A new platform for sensing urinary morphine based on carrier assisted electromembrane extraction followed by adsorptive stripping voltammetric detection on screen-printed electrode, Biosens. Bioelectron. 54 (2014) 189–194.

[167] A.R. Fakhari, M.H. Koruni, H. Ahmar, A. Shahsavani, S.K. Movahed, Electrochemical determination of dextromethorphan on reduced graphene oxide modified screen-printed electrode after electromembrane extraction, Electroanalysis 26 (2014) 1–9.

[168] J.P. Metters, R.O. Kadara, C.E. Banks, New directions in screen printed electroanalytical sensors: an overview of recent developments, Analyst 136 (2011) 1067–1076.

[169] H.M. Mohamed, Screen-printed disposable electrodes: Pharmaceutical applications and recent developments, Trends Anal. Chem. 82 (2016) 1–11.

[170] R.A.S. Couto, J.L.F.C. Lima, M.B. Quinaz, Recent developments, characteristics and potential applications of screen-printed electrodes in pharmaceutical and biological analysis, Talanta 146 (2016) 801–814.

[171] J. Hu, S.Q. Wang, L. Wang, F. Li, B. Pingguan-Murphy, T.J. Lu, F. Xu, Advances in paper-based point-of-care diagnostics, Biosens. Bioelectron. 54 (2014) 585–597.

[172] A. Abellán-Llobregat, I. Jeerapan, A. Bandodkar, L. Vidal, A. Canals, J. Wang, E. Morallón, A stretchable and screen-printed electrochemical sensor for glucose determination in human perspiration, Biosens. Bioelectron. 91 (2017) 885–891.

[173] M. Li, Y.-T. Li, D.-W. Li, Y.-T. Long, Recent developments and applications of screen-printed electrodes in environmental assays-A review, Anal. Chim. Acta 734 (2012) 31–44.

[174] J.M. Pingarrón, P. Sánchez, Química Electroanalítica. Fundamentos y Aplicaciones, 1a ed., Editorial Síntesis, Madrid (España), 2003.

Referencias

114

[175] J. Wang, Analytical Electrochemistry, 2nd ed., Wiley-VCH, New York (USA), 2001.

III. Parte experimental, resultados

y discusión

Capítulo 1

Parte experimental, resultados y discusión

121

Zeolite/iron oxide composite as sorbent for magnetic solid-phase extraction of benzene, toluene, ethylbenzene and xylenes

from water samples prior to gas chromatography-mass spectrometry

Elena Fernández, Lorena Vidal and Antonio Canals

Departamento de Química Analítica, Nutrición y Bromatología e Instituto Universitario de Materiales, Universidad de Alicante, P.O. Box 99, 03080 Alicante, Spain.

Received: 29 March 2016 / Received in revised form: 14 June 2016 / Accepted: 15 June 2016 / Available online: 16 June 2016

Abstract

This study reports a new composite based on ZSM-5 zeolite decorated with iron oxide magnetic nanoparticles as a valuable sorbent for magnetic solid-phase extraction (MSPE). A proposal is made to determine benzene, toluene, ethylbenzene and xylenes (BTEX) as model analytes in water samples using gas chromatography-mass spectrometry. A two-step multivariate optimization strategy, using Plackett-Burman and circumscribed central composite designs, was employed to optimize experimental parameters affecting MSPE. The method was evaluated under optimized extraction conditions (i.e., amount of sorbent, 138 mg; extraction time, 11 min; sample pH, pH of water (i.e., 5.5-6.5); eluent solvent volume, 0.5 mL; and elution time, 5 min), obtaining a linear response from 1 to 100 µg L-1 for benzene; from 10 to 100 µg L-1 for toluene, ethylbenzene and o-xylene; and from 10 to

Journal of Chromatograpy A 1458 (2016) 18-24

Capítulo 1

122

75 µg L-1 for m,p-xylene. The repeatability of the proposed method was evaluated at a 40 µg L-1 spiking level and coefficients of variation ranged between 8 and 11% (n=5). Limits of detection were found to be 0.3 µg L-1 for benzene and 3 µg L-1 for the other analytes. These values satisfy the current normative of the Environmental Protection Agency and European Union for BTEX content in waters for human consumption. Finally, drinking water, wastewater and river water were selected as real water samples to assess the applicability of the method. Relative recoveries varied between 85% and 114% showing negligible matrix effects.

Keywords: ZSM-5 zeolite, magnetic solid-phase extraction, gas chromatography-mass spectrometry, BTEX, water samples.

1. Introduction

Zeolites are porous crystalline aluminosilicates formed by SiO4 and AlO4 tetrahedra interconnected by corner oxygen atoms. Briefly, zeolites can be considered as pure silica frameworks with Si4+ substituted by Al3+. The presence of each AlO4 tetrahedron generates an overall negative charge, which is compensated by extra-framework cations (e.g., Na+, K+, Ca2+) placed within the pores and channels [1].

Zeolites exist as natural minerals (e.g., mordenite, clinoptilolite) and in synthetic forms (e.g., Y, ZSM-5, beta), and both types of zeolites have been used in industrial and commercial applications [2]. Nowadays, almost 200 different structures are known and accepted by the Structure Commission of the International Zeolite Association [3]. Both morphology (i.e., pores and channels size and shape) and chemical composition confer different properties to zeolites, which may be considered for specific applications. In general, frameworks with a high Si/Al ratio possess greater thermal stability, higher hydrophobicity and acidity, and lower ion-exchange capacity [1]. The fact that zeolites can be designed with different morphologies, chemical composition and properties has enabled their extensive application to different fields, mainly: (i) catalysis (e.g., cracking processes in petrochemical industry [4]); (ii) adsorption (e.g., purification of gas streams in petrochemical industry [4] or


123

the removal of organic compounds from contaminated waters [5]); and (iii) ion-exchange (e.g., water softening in detergents or heavy metal removal from contaminated waters [1,5]). Additionally, the ever growing interest in the synthesis of new zeolites has promoted innovative applications such as solar cell technology, microelectronics, medicine and holographic sensors [6,7].

Solid-phase extraction (SPE) is one of the most widely used sample preparation techniques providing analyte isolation, preconcentration and sample clean-up [8]. Typically, SPE consists of cartridges packed with sorbent where the analyte is retained when liquid samples flow through it. Then, a proper solvent is employed to elute and recover the analyte for further determination [8]. Several sorbents with different properties have been employed in SPE, including inorganic oxides, porous polymers, molecularly imprinted polymers, biosorbents or nanomaterials [8-10]. In addition, some publications have reported the use of zeolites for the preconcentration of inorganic [11-13] and organic [14,15] analytes. The original SPE has been modified several times to date, mainly related to miniaturization or automation [16]. Zeolites have also been proposed as sorbents in new SPE protocols [17-21]. Recently, magnetic solid-phase extraction (MSPE) has received great interest since it reduces sample preparation time and facilitates sorbent manipulation [16,22]. In MSPE, the magnetic sorbent is dispersed into the aqueous phase and after extraction, it is easily separated from the sample solution by applying an external magnetic field, and thus avoids time-consuming filtration or centrifugation steps for phase separation. Next, target analytes can be desorbed using a proper eluent solvent or temperature for further determination [16,22]. A recent publication has reported the use of zeolite/Fe3O4 composite as new sorbent for MSPE for the first time [23]. In this work, phthalates are determined at trace levels in aqueous samples after extraction with clinoptilolite zeolite loaded on Fe3O4 nanoparticles [23].

Benzene, toluene, ethylbenzene and xylene (BTEX) are compounds of public health concern due to their toxic and carcinogenic properties [24]. BTEX are extensively used in many industrial processes and benzene, toluene and ethylbenzene are considered priority pollutants by the Environmental Protection Agency (EPA) [24]. Therefore, their presence in the environment

Capítulo 1

124

must be controlled. Numerous methods including SPE techniques have been proposed for BTEX determination in water samples [25-30].

The purpose of our work is to develop an analytical method in which a composite based on ZSM-5 zeolite and iron oxide magnetic nanoparticles is presented as a valuable new sorbent for MSPE. BTEX determination in water samples has been proposed as a model analytical application using gas chromatography-mass spectrometry (GC-MS) for analysis. To the best of our knowledge, this is the first time that ZSM-5/iron oxide magnetic composite has been employed for BTEX extraction and preconcentration. Various parameters of the proposed MSPE have been optimized by the use of a multivariate optimization strategy. Finally, the applicability of the proposed method to determine BTEX at trace levels in real water samples has been evaluated.

2. Experimental part

2.1. Reagents and water samples

Benzene, ethylbenzene and xylene isomers of analytical standard grade were obtained from Fluka (Steinheim, Germany), whereas toluene LC grade (≥99.9%) was bought from Scharlau Chemie (Barcelona, Spain). Mixed stock standard solution of BTEX was prepared at a concentration of 1000 µg L-1 in acetone LC grade from Sigma-Aldrich (Steinheim, Germany) and stored in the dark at 4 ºC. Working solutions (1-100 µg L-1) were prepared by diluting stock solution with acetone. Aqueous working solutions were daily prepared in ultrapure water (resistivity of 18.2 MΩ cm at 25 ºC) from a Millipore Direct System Q5TM purification system from Ibérica S.A. (Madrid, Spain). H3PO4 (85%) from Scharlau Chemie and KH2PO4 and K2HPO4 pro-analysis from Merck (Darmstadt, Germany) were employed for preparing buffer solutions during sample pH optimization. FeCl3·6H2O and FeSO4·7H2O reactive grade were obtained from Sigma-Aldrich and NaOH (97%, pellets) from Scharlau Chemie.

ZSM-5 zeolite (CBV 3024E, SiO2/Al2O3 mole ratio=30) in the ammonium nominal cation form was purchased from Zeolist International (Conshohocken, PA, USA).


125

As real water samples, we used drinking water from a drinking-water treatment plant in Albacete (Spain), industrial wastewater from Zaragoza (Spain) and river water from Murcia (Spain). Samples were collected in amber glass containers and stored in the dark at 4 ºC. Drinking and river waters were used without any further pretreatment but wastewater was filtered with a 0.45 µm pore-size nylon filter before use. Water samples were analyzed under optimized conditions of the proposed method, and target analytes were not detected.

2.2. Instrumentation

X-ray photoelectron spectroscopy (XPS) was used for the chemical characterization of iron oxide in the ZSM-5/iron oxide composite. XPS measurements were performed with an automatic K-Alpha spectrometer from Thermo-Scientific (Waltham, MA, USA). The spectra were recorded with pass energy of 50 eV, X-ray spot size of 400 µm, and step size of 0.1 eV.

Surface area (BET) and pore volume (t-plot method) of ZSM-5 zeolite and ZSM-5/iron oxide composite were measured by nitrogen adsorption at 77 K using an Autosorb-6 analyzer from Quantachrome Instruments (Hook, UK).

Two chromatographic systems were employed for MSPE optimization, and method validation and real samples analysis, respectively. For MSPE optimization, chromatographic analysis were performed on a gas chromatograph (model 7890A) from Agilent Technologies (Santa Clara, CA, USA) equipped with a splitless/split automatic injector and a flame ionization detector (FID). A capillary column HP-5 (5% phenyl-95% dimethylpolysiloxane, 30 m × 0.32 mm I.D., 0.25 µm film thickness) was from J&W Scientific (Folsom, CA, USA). The injector temperature was maintained at 200 ºC and the injection volume was 1.0 µL in the split mode (split ratio 1:10). The oven temperature program was initially set at 40 ºC and was raised by 6 ºC min-1 up to 100 ºC (held 5 min). Helium (99.999%) from Air Liquide (Madrid, Spain) was used as the carrier gas at a constant flow rate of 1 mL min-

1. The detector temperature was set at 300 ºC. Method validation and real samples analysis were carried out using GC-

MS. The gas chromatograph (model 6890N) was from Agilent Technologies

Capítulo 1

126

and the capillary column TRB-624 (6% cyanopropylphenyl-94% dimethylpolysiloxane, 30 m × 0.25 mm I.D., 1.4 µm film thickness) was from Teknokroma (Barcelona, Spain). Samples (2 µL) were injected in the split mode (split ratio 1:5) with the injector maintained at 250 ºC. The injection liner (19251-60540 straight split) was from Agilent Technologies. The oven temperature was initially set at 35 ºC for 5 min, followed by a ramp of 5 ºC min-1 up to 150 ºC. Helium (99.999%) was used as the carrier gas at a constant flow rate of 1 mL min-1. The MS detector (model 5973N) was from Agilent Technologies. Electron impact ionization was used with ionization energy of 70 eV. The mass source and quadrupole were set at 250 and 150 ºC, respectively. Measurements were taken with a solvent delay of 7 min and in selected ion monitoring (SIM) mode at the following mass/charge ratios: 77 and 78 from minute 7-11 for benzene determination; 91 and 92 from minute 11-16 for toluene; and 91 and 106 from minute 16-28 for ethylbenzene and xylenes determination. Figure 1 shows typical chromatograms after MSPE of a blank and a standard solution spiked at 100 µg L-1 with target analytes. It should be mentioned that the chromatographic separation of m-xylene and p-xylene isomers was not performed under the above conditions.


127

Figure 1. Total ion chromatograms obtained in the SIM mode after MSPE of a blank (green line) and a standard solution spiked at 100 µg L-1 with target analytes (black line).

Abu

ndan

ce u

nits

(au)

Time (min)

Benzene

Toluene

Ethylbenzene

m,p-xylene

o-xylene

Capítulo 1

128

2.3. Synthesis of ZSM-5/iron oxide composite

The composite was synthesized considering a ZSM-5/iron oxide ratio 3/1 (w:w), being a compromise value between iron oxide load (proven to decrease the available surface area and micropore volume [31]) and the easy manipulation of the composite under a magnetic field. Thereby, the composite was prepared from a suspension of 1 g of ZSM-5 zeolite in 250 mL of water solution with FeCl3·6H2O (0.778 g) and FeSO4·7H2O (0.400 g). After homogeneous mixing, 5 mL of NaOH 5 M was added drop wise to precipitate the iron oxide. The mixture was then stirred for 2 h at room temperature. The resulting composite was cleaned with ultrapure water until clear washing waters were obtained, using a neodymium magnet to separate magnetic from non-magnetic material. Finally, the composite was dried at 100 ºC overnight. Figure S1 shows a photograph of the magnetic attraction between the synthesized composite and the magnet. Before MSPE, the composite was kept at 200 ºC for 2 h in order to remove the excess of physisorbed water and other residual compounds.

Figure S1. Photograph showing the magnetic attraction between ZSM-5/iron oxide composite and neodymium magnet.


129

2.4. Magnetic solid-phase extraction

Firstly, 138 mg of the magnetic composite were placed in a 22 mL glass vial. Then, 20 mL of aqueous standard or sample solution were added and the mixture was shaken vigorously for 11 min. After extraction, the neodymium magnet was used to attract the composite to the bottom of the vial and, thus, the water phase was easily removed using a glass pipette. Then, elution was carried out using 0.5 mL of acetone and shaking gently for 5 min in order to prevent sorbent sticking in the vial walls. Acetone was chosen as eluent since previous results (data not shown) revealed it performed better than other organic solvents (e.g., acetonitrile, methanol). Finally, the organic phase was separated from the magnetic composite using the neodymium magnet, retrieved with a syringe and filtered with 0.45 µm pore-size nylon filters for injection in GC-FID (for MSPE optimization studies) or GC-MS (for method validation and real sample analysis).

2.5. Data processing

A two-step multivariate optimization strategy, using Plackett-Burman and circumscribed central composite designs, was carried out to determine the optimum conditions for MSPE. The statistical software NEMRODW® (“New Efficient Methodology for Research using Optimal Design”) from LPRAI (Marseille, France) was used to build the experimental design matrices and evaluate the results. Peak areas of benzene, toluene and ethylbenzene obtained with GC-FID were individually used as response functions for optimization. It should be mentioned that xylenes could not be included in optimization studies since the lower sensitivity of the detector used in this step (i.e., GC-FID) did not allow their quantification.

Capítulo 1

130

3. Results and discussion

3.1. Characterization of ZSM-5/iron oxide composite

Figure 2 shows a high resolution XPS spectrum in the Fe 2p region. Comparing the spectrum obtained with database spectra [32], it can be concluded that iron is in its highest oxidation state (Fe3+) forming Fe2O3. The results suggest that, although Fe2+ and Fe3+ salts (i.e., FeSO4·7H2O and FeCl3·6H2O, respectively) were employed to synthesize iron oxide nanoparticles, Fe2+ was oxidized to Fe3+ during composite preparation and Fe2O3 was finally obtained instead of the mixed oxide (i.e., Fe3O4) [31]. According to bibliography [33], both van der Waals and electrical forces are involved in the interaction between zeolites and Fe2O3 particles obtained by chemical precipitation.

Figure 2. High resolution XPS spectrum in the Fe 2p region for ZSM-5/iron oxide composite.

0.0E+00

4.0E+03

8.0E+03

1.2E+04

1.6E+04

2.0E+04

700705710715720725730735740745

Inte

nsity

(cps

)

Binding Energy (eV)

raw datafitting peaks

Fe 2p1/2

Fe 2p3/2

satellite


131

For pure ZSM-5 zeolite and ZSM-5/iron oxide composite, BET surface areas of 442 m2 g-1 (Vmicropores=0.161 cm3 g-1) and 292 m2 g-1 (Vmicropores=0.073 cm3 g-1), were respectively obtained from nitrogen adsorption isotherms (Figure S2) and BET theory. It is interesting to note the decrease in BET surface area and micropore volume in the magnetic composite, due to the presence of iron oxide nanoparticles [31]. Additionally, Figure S2 shows that adsorption at high pressures was greater for ZSM-5/iron oxide composite which could be related with the creation of mesopores on zeolite surface due to poorly dispersed iron oxide nanoparticles. Despite the decrease in surface area of ZSM-5/iron oxide compared with pure zeolite, the use of the magnetic composite as sorbent provides unique advantages due to its easy manipulation and shorter extraction times. In addition, the proposed method provides limits of detection and quantification low enough to satisfy current normative [34,35] (see Section 3.4). Finally, it is important to point out that extractions using Fe2O3 as extractant phase (without zeolite) were carried out and analytes were not detected in the eluates, showing that Fe2O3 did not possess a sorption capacity for target analytes.

Figure S2. Nitrogen adsorption-desorption isotherms for pure ZSM-5 zeolite and ZSM-5/iron oxide composite.

0

50

100

150

200

250

300

350

0.0 0.2 0.4 0.6 0.8 1.0

Ads

orbe

d vo

lum

e (c

m3 g-1

)

P/P0

ZSM-5/iron oxide

ZSM-5

Capítulo 1

132

3.2. Multivariate optimization

3.2.1. Screening

Screening experiments initially include many factors in order to identify which ones have important effects [36]. Fractional factorial designs are the most widely used in screening experiments. Plackett-Burman is a two-level fractional factorial design for studying up to k=N-1 variables in N runs, where N is a multiple of 4 [36]. Plackett-Burman design assumes that interaction between factors can be ignored so the main effects can be calculated with a reduced number of experiments, thereby saving time and resources. A Plackett-Burman design was used to construct the matrix of experiments, including five factors studied in eight runs. The five experimental factors selected at two levels were: amount of sorbent, sample pH, extraction time, eluent solvent volume and elution time. Table 1 shows the experimental factors and levels considered in the Plackett-Burman design whereas Table S1 shows the matrix of experiments. The eight experiments were randomly performed using 20 mL of aqueous standards with 20 mg L-1 of benzene, toluene and ethylbenzene due to preliminary investigations (about the extraction) at higher concentrations than 20 mg L-1 revealed that sorbent capacity was not saturated. The analysis was carried out using GC-FID.

Table 1. Experimental factors and levels of the Plackett-Burman design.

Factors Level

Low (-1) High (+1)

Amount of sorbent (mg) 50 100

Sample pH 3 9

Extraction time (min) 3 6

Eluent solvent volume (mL) 0.5 1 Elution time (min) 5 10


133

Table S1. Matrix of experiments of the Plackett-Burman design.

Run Amount of sorbent (mg)

Sample pH

Extraction time (min)

Eluent solvent volume (mL)

Elution time (min)

1 100 9 6 0.5 10

2 50 9 6 1 5

3 50 3 6 1 10

4 100 3 3 1 10

5 50 9 3 0.5 10

6 100 3 6 0.5 5

7 100 9 3 1 5

8 50 3 3 0.5 5

The data obtained were evaluated by ANOVA and the results were visualized with the Pareto charts shown in Figure S3. The length of each bar was proportional to the influence of the corresponding factor while the effects that exceed reference vertical lines can be considered significant with 95% of probability. In addition, negative and positive signs reveal whether the system response decreases or increases, respectively, when passing from the lowest to the highest level of the corresponding factor.

Capítulo 1

134

Figure S3 Pareto charts of the Plackett-Burman design obtained for: (a) benzene; (b) toluene; and (c) ethylbenzene.

(a)

(b)

(c)

Amount of sorbent (mg)

Sample pH



Elution time (min)


Sample pH



Elution time (min)


Sample pH



Elution time (min)


135

As shown in Figure S3, the amount of sorbent, extraction time and eluent solvent volume were the most important factors, having the same sign for the three studied analytes and, therefore, showing analogous behaviors during extraction. Although these factors were not significant in all cases, they were shown to have the greatest effects and were selected for the next optimization step. The amount of sorbent and extraction time, both with positive effects, were studied in the circumscribed central composite design (CCCD) whereas finally eluent solvent volume could not be considered. The eluent solvent volume showed a negative effect, which is easily explained considering that when one employs a smaller volume of eluent solvent one obtains a higher analyte concentration in the extract. However, volumes lower than 0.5 mL could not be easily handled after elution and, therefore, a volume of 0.5 mL of acetone was finally fixed for subsequent experiments. Sample pH and elution time showed non-significant effects and they were fixed at the most convenient experimental levels, without adjusting water pH, and with a 5 min elution time.

3.2.2. Optimization

Circumscribed central composite design (CCCD) was employed in this optimization step. This design combines a two-level full factorial design (2k) with 2k star points, where k is the number of factors being optimized, and one point at the center of the experimental region, which can be run n times. In

order to ensure model rotatability, star points were set at α=√2k4= 1.41 whereas

the central point was repeated three times to provide an orthogonal design [36]. CCCD was used to evaluate the main effects, interaction effects and quadratic effects of the amount of sorbent and extraction time. Table 2 shows the low, central and high levels, and the star points of the considered factors.

Capítulo 1

136

Table 2. Factors, low, central and high levels, and star points used in CCCD.

Factors Level Star points

(1.41) Low (-1)

Central (0)

High (+1) - +

Amount of sorbent (mg) 100 150 200 79 221

Extraction time (min) 4 7 10 3 11

Eleven experiments (Table S2) were randomly carried out using 20 mL of aqueous standards with 20 mg L-1 of benzene, toluene and ethylbenzene. GC-FID was used for the analysis.

Table S2. Matrix of experiments of CCCD.

Run Amount of sorbent (mg) Extraction time (min)

1 100 4

2 200 4

3 100 10

4 200 10

5 79 7

6 221 7

7 150 3

8 150 11

9 150 7

10 150 7

11 150 7


137

Data obtained were also evaluated by ANOVA. The quadratic regression model was significant in all cases (data not shown). Coefficients of variation (CV) at central point (n=3) were 6%, 7% and 11% for benzene, toluene and ethylbenzene, respectively. Response surfaces of the CCCD are shown in Figure 3. As can be observed, the higher the extraction time, the bigger the signals; an optimum value of 11 min was set for all analytes under study. Nevertheless, the influence of the extraction time was less important in benzene, probably because this analyte is easier to extract due to steric effects and size. Furthermore, an increase amount of sorbent led to an initial increase in signals, followed by a decrease. This could be related with nanoparticle aggregation turning into: (i) lower surface area available for extraction; and/or, (ii) less desorption efficiency in the elution step. Optimum values for the amount of sorbent were 127, 142 and 144 mg for benzene, toluene and ethylbenzene, respectively. A compromise value (i.e., mean value) of 138 mg was adopted to validate the method.

Figure 3. Response surfaces of CCCD obtained by plotting the amount of sorbent vs. extraction time for: (a) benzene; (b) toluene; and (c) ethylbenzene.

Extraction time

Amount of sorbent

(a) Peak area (pA)

Capítulo 1

138

Figure 3 (cont.). Response surfaces of CCCD obtained by plotting the amount of sorbent vs. extraction time for: (a) benzene; (b) toluene; and (c) ethylbenzene.

(b) Peak area (pA)

Extraction time

Amount of sorbent

(c) Peak area (pA)

Extraction time

Amount of sorbent


139

Based on the results shown above, the MSPE experimental conditions selected for BTEX extraction were: amount of sorbent, 138 mg; extraction time, 11 min; sample pH, water pH without adjustment; eluent solvent volume, 0.5 mL; and elution time, 5 min.

3.3. Sorbent reutilization study

The possibility of reusing ZSM-5/iron oxide composite was studied considering the environmental importance of saving chemicals and wastes. Twelve consecutive extractions were conducted using the same ZSM-5/iron oxide composite, which was dried at 200 ºC for 2 h after each extraction. The experiments were carried out in triplicate, employing 10 mg L-1 of benzene, toluene and ethylbenzene aqueous standard solutions and GC-FID as detector. As mentioned before, xylenes were not included in this study since the lower sensitivity of GC-FID did not allow their quantification. As shown in Figure 4, signals remained practically constant during the twelve experiments with CV values ranged between 2 and 15% and, thus demonstrating the possibility of reusing ZSM-5/iron oxide composite at least twelve times.

Figure 4. Study of sorbent reutilization using the same ZSM-5/iron oxide composite in twelve consecutive extractions.

0102030405060708090

100

1 2 3 4 5 6 7 8 9 10 11 12

Peak

are

a (p

A)

Extraction number

BenzeneTolueneEthylbenzene

Capítulo 1

140

3.4. Analytical figures of merit

Quality parameters of the proposed method were assessed under MSPE optimized conditions. The concentration range studied was from 1 to 100 µg L-

1 for all the analytes and the final linear working ranges are shown in Table 3. Other main analytical parameters of the proposed method are also summarized in Table 3. The lower concentrations of linear ranges were limited by the limit of quantification (LOQ). The upper end for m,p-xylene was established at 75 µg L-1 since standards of 75 µg L-1 and 100 µg L-1 provided the same response. Correlation coefficients (r) ranged from 0.950 to 0.997 (Table 3). Student’s t-test [37] was applied to assess linearity, obtaining the following t calculated values: 16.3 (r=0.993; N=6) for benzene; 14.5 (r=0.991; N=6) for toluene; 13.3 (r=0.990; N=6) for ethylbenzene; 20.9 (r=0.997; N=5) for m,p-xylene; and 5.84 (r=0.950; N=6) for o-xylene. In all cases, the null hypothesis of no correlation for a 5% significance level and 3 or 4 degrees of freedom (t0.05,3=3.18 and t0.05,4=2.78, respectively) could be rejected and we concluded that a significant correlation does exist. The sensitivity of the instrumental measurements estimated by the slope of the calibration curves ranged between (24800±1700) au µg-1 L for toluene and (180±30) au µg-1 L for o-xylene. The repeatability of the proposed method, expressed as CV, was evaluated by analyzing five aqueous standards at BTEX concentration of 40 µg L-1. CV values ranged between 8 and 11% (Table 3). Enrichment factors (EFs) were evaluated through the slope ratio of calibration curves with and without MSPE. As can be seen in Table 3, EFs were very similar for benzene, toluene and ethylbenzene. However, xylene isomers gave lower extraction performance, with the EF value below 1 for o-xylene. This means that there was no preconcentration for this analyte, probably due to sterically hindered extraction. These EFs values agree with the efficiency for the complete sample preparation procedure which was 19% for benzene, 17% for toluene, 16% for ethylbenzene, 5% for m,p-xylene and 2% for o-xylene. Limit of detection (LOD) and LOQ were estimated according to 3Sb and 10Sb criteria, respectively, where Sb is the standard deviation of the blank [37]. The obtained values (Table 3) satisfy current normative according to EPA [34] and the European Union [35] for BTEX content in waters intended for human consumption. EPA establishes maximum contamination levels of 5 µg L-1 for benzene, 1 mg L-1 for toluene,


141

0.7 mg L-1 for ethylbenzene and 10 mg L-1 for xylenes [34]. The European Union is more restrictive with levels of benzene in drinking waters due to its carcinogenic properties, and establishes a threshold of 1 µg L-1. By contrast, toluene, ethylbenzene and xylenes are not included in European legislation [35]. Finally, we want to point out that LODs obtained for benzene and toluene are lower than those previously reported (i.e., 11 µg L-1 and 13 µg L-1, respectively) for a method in which ZSM-5 zeolite was employed as sorbent for headspace SPE [21].

Table 3. Main analytical parameters of the proposed method (MSPE-GC-MS).

Analyte Working range (µg L-1) ra CVb

(%) LODc

(µg L-1) LOQd

(µg L-1) EFe

Benzene 1-100 0.993 (6) 9 0.3 1 7.6 Toluene 10-100 0.991 (6) 8 3 10 7.0

Ethylbenzene 10-100 0.990 (6) 9 3 10 6.5 m,p-xylene 10-75 0.997 (5) 9 3 10 1.7

o-xylene 10-100 0.950 (6) 11 3 10 0.6 a Correlation coefficient: number of calibration points in parentheses. b Coefficient of variation: mean value for five replicate analyses of a 40 µg L-1 spiked solution. c Limit of detection: calculated using blank signal plus three times its standard deviation. d Limit of quantification: calculated using blank signal plus ten times its standard deviation. e Enrichment factor: calculated as slope ratio between calibration curves with and without MSPE.

3.5. Analysis of real water samples

The applicability of the proposed method to determine BTEX in real water samples was evaluated. Three water samples (namely drinking water, wastewater and river water) were employed to assess matrix effects using recovery studies. It should be noted that in previous analyses with the proposed method, none of the selected water samples had initial detectable BTEX concentrations. Three replicated analyses of each water sample were carried out at 40 µg L-1 spiking level. Table 4 shows the relative recoveries determined as the ratio of the signals found after MSPE in real and ultrapure water samples spiked at the same concentration level.

Capítulo 1

142

Table 4. Relative recoveries and CV values (in parentheses) obtained for the target analytes in the three studied real water samples.

Analyte Relative recoveries (%) and CV values

in parentheses (%)a

Drinking water Wastewater River water

Benzene 94 (4) 112 (8) 112 (11) Toluene 92 (6) 112 (11) 113 (2)

Ethylbenzene 85 (3) 113 (8) 111 (3) m,p-xylene 93 (9) 98 (11) 114 (4)

o-xylene 87 (8) 93 (2) 112 (8) a Three replicate analysis at 40 µg L-1 spiking level.

Results showed relative recoveries varying from 85 to 94% for drinking water, between 93 and 113% for wastewater, and between 111 and 114% for river water. In all cases, CV values were below 11%. According to these results, we can conclude that matrix effects were not significant for the determination of BTEX in the three selected water samples.

4. Conclusions

A new MSPE-GC-MS procedure using ZSM-5/iron oxide composite as sorbent has been proposed in this work. The magnetic composite presents remarkable advantages such as low cost, rapid and simple synthesis, easy manipulation under a magnetic field and reuse options. To our knowledge, this is the first time that ZSM-5/iron oxide composite has been employed for BTEX extraction and preconcentration from water samples. Good extraction efficiencies were obtained for benzene, toluene and ethylbenzene. However, EFs were lower for xylenes, and preconcentration for o-xylene isomer was not obtained, probably due to steric effects. Nevertheless, the results showed LODs that satisfy the current EPA and European Union normative for BTEX content in waters for human consumption, especially for benzene, which presents the most restrictive levels (i.e., 5 and 1 µg L-1, respectively). In addition, LODs were lower than those previously reported for benzene and toluene in a method


143

using ZSM-5 zeolite as sorbent for headspace SPE. Finally our study demonstrates the ability of the proposed method to determine BTEX at trace levels in real water samples.

Zeolites possess unique sorbent properties, such as ordered crystalline structure with well-defined pore size and shape. Therefore, zeolites can act as molecular sieves, which are very useful for selective extractions. This work can be considered as a starting point for the use of ZSM-5/iron oxide composite in MSPE. In future works, the application of ZSM-5/iron oxide composite can be extended to different target analytes (i.e., organic and inorganic) and samples (e.g., environmental, food and biological).

Acknowledgements

The authors would like to thank the Spanish Ministry of Science and Innovation (project no. CTQ2011-23968), Generalitat Valenciana (Spain) (projects nos. GVA/2014/096 and PROMETEO/2013/038) for the financial support. E. Fernández also thanks Ministry of Education for her FPU grant (FPU13/03125). The authors would also like to thank C. Cortada from iNTERLAB (Alicante, Spain) for the real water samples supply and F. Barraclough for the English revision.

References

[1] S. Kulprathipanja, Zeolites in Industrial Separation and Catalysis, 1st ed., Wiley-VCH Weinheim, Germany, 2010.

[2] C. Martínez, J. Pérez-Pariente, Zeolites and Ordered Mesoporous Solids: Fundamentals and Applications, 1st ed., Universitat Politècnica de València, Valencia (Spain), 2011.

[3] Structure Commision of the International Zeolite Association (January, 2016) http://www.iza-structure.org/.

[4] W. Vermeiren, J.-P. Gilson, Impact of zeolites on the petroleum and petrochemical industry, Top. Catal. 52 (2009) 1131-1161.

[5] S. Wang, Y. Peng, Natural zeolites as effective adsorbents in water and wastewater treatment, Chem. Eng. J. 156 (2010) 11-24.

Capítulo 1

144

[6] M. Zaarour, B. Dong, I. Naydenova, R. Retoux, S. Mintova, Progress in zeolite synthesis promotes advanced applications, Microporous Mesoporous Mat. 189 (2014) 11-21.

[7] J. Li, A. Corma, J. Yu, Synthesis of new zeolite structures, Chem. Soc. Rev. 44 (2015) 7112-7127.

[8] C.F. Poole, New trends in solid-phase extraction, TrAC Trends Anal. Chem. 22 (2003) 362-373.

[9] L. Okenicová, M. Zemberyová, S. Procházková, Biosorbents for solid-phase extraction of toxic elements in waters, Environ. Chem. Lett. 14 (2016) 67-77.

[10] B. Hu, M. He, B. Chen, Nanometer-sized materials for solid-phase extraction of trace elements, Anal. Bioanal. Chem. 407 (2015) 2685-2710.

[11] A. Mostafavi, D. Afzali, M.A. Taher, Atomic absorption spectrometric determination of trace amounts of copper and zinc after simultaneous solid-phase extraction and preconcentration onto modified natrolite zeolite, Anal. Sci. 22 (2006) 849-853.

[12] A. Malekpour, S. Hajialigol, M.A. Taher, Study on solid-phase extraction and flame atomic absorption spectrometry for the selective determination of cadmium in water and plant samples with modified clinoptilolite, J. Hazard. Mater. 172 (2009) 229-233.

[13] A.S. Saljooghi, Z.S. Saljoghi, Natural analcime zeolite modified with 2,3,5,6-tetra(2-pyridyl)pyrazine for preconcentration and determination of trace amounts of cadmium by flame atomic absorption spectrometry, Toxicol. Ind. Health 28 (2012) 771-778.

[14] Y.S. Al-Degs, A.H. El-Sheikh, M.A. Al-Ghouti, B. Hemmateenejad, G.M. Walker, Solid-phase extraction and simultaneous determination of trace amounts of sulphonated and azo sulphonated dyes using microemulsion-modified-zeolite and multivariate calibration, Talanta 75 (2008) 904-915.

[15] P. Arnnok, N. Patdhanagul, R. Burakham, An on-line admicellar SPE-HPLC system using CTAB-modified zeolite NaY as sorbent for determination of carbamate pesticides in water, Chromatographia 78 (2015) 1327-1337.


145

[16] J. Płotka-Wasylka, N. Szczepanska, M. de la Guardia, J. Namiesnik, Miniaturized solid-phase extraction techniques, TrAC Trends Anal. Chem. 73 (2015) 19-38.

[17] A.A. Matin, R. Maleki, M.A. Farajzadeh, K. Farhadi, R. Hosseinzadeh, A. Jouyban, Headspace SPME-GC method for acetone analysis and its biomedical application, Chromatographia 66 (2007) 383-387.

[18] T.P. Lee, B. Saad, E.P. Ng, B. Salleh, Zeolite Linde Type L as micro-solid phase extraction sorbent for the high performance liquid chromatography determination of ochratoxin A in coffee and cereal, J. Chromatogr. A 1237 (2012) 46-54.

[19] W.B. Wilson, A.A. Costa, H. Wang, J.A. Dias, S.C.L. Dias, A.D. Campiglia, Analytical evaluation of BEA zeolite for the pre-concentration of polycyclic aromatic hydrocarbons and their subsequent chromatographic analysis in water samples, Anal. Chim. Acta 733 (2012) 103-109.

[20] W.B. Wilson, A.A. Costa, H. Wang, A.D. Campiglia, J.A. Dias, S.C.L. Dias, Pre-concentration of water samples with BEA zeolite for the direct determination of polycyclic aromatic hydrocarbons with laser-excited time-resolved Shpol’skii spectroscopy, Microchem. J. 110 (2013) 246-255.

[21] S. Goda, R. Selyanchyn, T. Nozoe, H. Matsui, S.W. Lee, Development of a thin-film microextraction device based on ZSM-5/Tenax TA for VOC detection in liquid samples, J. Anal. Bioanal. Tech. S12 (2014) 004/1-004/7.

[22] M. Wierucka, M. Biziuk, Application of magnetic nanoparticles for magnetic solid-phase extraction in preparing biological, environmental and food samples, TrAC Trends Anal. Chem. 59 (2014) 50-58.

[23] A. Mollahosseini, M. Toghroli, M. Kamankesh, Zeolite/Fe3O4 as a new sorbent in magnetic solid-phase extraction followed by gas chromatography for determining phthalates in aqueous samples, J. Sep. Sci. 38 (2015) 3750-3757.

[24] Enviromental Protection Agency (January, 2016), http://www.epa.gov/eg/toxic-and-priority-pollutants-under-clean-water-act.

Capítulo 1

146

[25] A. Sarafraz-Yazdi, H. Piri moghadam, Z. Es’haghi, S. Sepehr, Comparative study of the three sol-gel based solid phase microextraction fibers in extraction of BTEX from water samples using gas chromatography-flame ionization detection, Anal. Methods 2 (2010) 746-752.

[26] A. Sarafraz-Yazdi, A. Amiri, G. Rounaghi, H.E. Hosseini, A novel solid-phase microextraction using coated fiber based sol-gel technique using poly(ethylene glycol) grafted multi-walled carbon nanotubes for determination of benzene toluene, ethylbenzene and o-xylene in water samples with gas chromatography-flame ionization detector, J. Chromatogr. A 1218 (2011) 5757-5764.

[27] J.N. Bianchin, G. Nardini, J. Merib, A.N. Dias, E. Martendal, E. Carasek, Simultaneous determination of polycyclic aromatic hydrocarbons and benzene toluene, ethylbenzene and xylene in water samples using a new sampling strategy combining different extraction modes and temperatures in a single extraction solid-phase microextraction-gas chromatography-mass spectrometry procedure, J. Chromatogr. A 1233 (2012) 22-29.

[28] A. Sarafraz-Yazdi, H. Vatani, A solid phase microextraction coating based on ionic liquid sol-gel technique for determination of benzene toluene, ethylbenzene and o-xylene in water samples using gas chromatography flame ionization detector, J. Chromatogr. A 1300 (2013) 104-111.

[29] A. Sarafraz-Yazdi, A. Yekkebashi, A non-ionic surfactant-mediated sol-gel coating for solid-phase microextraction of benzene, toluene, ethylbenzene and o-xylene in water samples using a gas chromatography-flame ionization detector, New J. Chem. 38 (2014) 4486-4493.

[30] A. Sarafraz-Yazdi, A. Zendegi-Shiraz, Z. Es’haghi, M. Hassanzadeh-Khayyat, Synthesis and characterization of composite polymer polyethylene glycol grafted flower-like cupric nano oxide for solid phase microextraction of ultra-trace levels of benzene, toluene, ethyl benzene and o-xylene in human hair and water samples, J. Chromatogr. A 1418 (2015) 21-28.


147

[31] L.C.A. Oliveira, D.I. Petkowicz, A. Smaniotto, S.B.C. Pergher, Magnetic zeolites: a new adsorbent for removal of metallic contaminants from water, Water Res. 38 (2004) 3699-3704.

[32] Thermo Scientific database (January, 2016), www.lasurface.com. [33] D. Feng, C. Aldrich, H. Tan, Removal of heavy metal ions by carrier

magnetic separation of adsorptive particulates, Hydrometallurgy 56 (2000) 359-368.

[34] Environmental Protection Agency (January, 2016), http://www.epa.gov/ dwstandardsregulations.

[35] COUNCIL DIRECTIVE 98/83/EC of 3 November on the quality of water intended for human consumption (January, 2016), http://ec.europa.eu/ environment/water/water-drink/legislation en.html.

[36] D.C. Montgomery, Design and Analysis of Experiments, 7th ed., Wiley New Jersey, USA, 2009.

[37] J.N. Miller, J.C. Miller, Statistics and Chemometrics for Analytical Chemistry, 5th ed., Pearson Prentice Hall London, UK, 2005.

Capítulo 2


151

Hydrophilic magnetic ionic liquid for headspace single-drop microextraction of chlorobenzenes prior to thermal desorption

gas chromatography-mass spectrometry


Departamento de Química Analítica, Nutrición y Bromatología e Instituto Universitario de Materiales, Universidad de Alicante, P.O. Box 99, E-03080 Alicante, Spain.

Abstract

A new, fast, easy to handle and environmentally friendly headspace single-drop microextraction (HS-SDME) based on a magnetic ionic liquid (MIL) as extractant solvent is presented. A small drop of the MIL 1-ethyl-3-methylimidazolium tetraisothiocyanatocobaltate (II) ([Emin]2[Co(NCS)4]) is located on one end of a small Nd magnet to extract nine chlorobenzenes (1,2-dichlorobenzene, 1,3-dichlorobenzene, 1,4-dichlorobenzene, 1,2,3-trichlorobenzene, 1,2,4-trichlorobenzene, 1,3,5-trichlorobenzene, 1,2,3,4-tetrachlorobenzene, 1,2,4,5-tetrachlorobenzene, and pentachlorobenzene) as model analytes from water samples prior to thermal desorption gas chromatography-mass spectrometry determination. A multivariate optimization strategy was employed to optimize experimental parameters affecting HS-SDME. The method was evaluated under optimized extraction conditions (i.e., sample volume, 20 mL; MIL volume, 1 µL; extraction time, 10 min; stirring speed, 1500 rpm; and ionic strength, 15% NaCl (w/v)), obtaining a linear response from 0.05 to 5 µg L-1 for all analytes. The repeatability of the proposed method was evaluated at 0.7 and 3 µg L-1 spiking levels and coefficients of variation ranged between 3 and 18% (n=3). Limits of detection were in the order of ng L-1 ranging from 3.7 ng L-1 for 1,4-dichlorobenzene to 7.6 ng L-1 for 1,2,4,5-tetrachlorobenzene. Finally, tap water, pond water and wastewater were selected as real water samples to assess the applicability of the method. Relative recoveries varied between 82 and 114% showing negligible matrix effects.

Capítulo 2

152

Keywords: magnetic ionic liquid, headspace single-drop microextraction, thermal desorption gas chromatography-mass spectrometry, chlorobenzenes, water samples.

1. Introduction

Ionic liquids (ILs) are a group of organic salts with melting points at or below 100 ºC. ILs possess versatile and relevant physico-chemical properties (namely, low vapor pressure, good chemical and thermal stability, nonflammability, high ionic conductivity, wide electrochemical potential window, and good extractability for organic compounds and metal ions), being responsible of the vast number of applications in analytical chemistry, including liquid-phase microextraction (LPME) techniques [1]. The use of ILs in these techniques not only has helped to overcome problems associated with the use of organic solvents (i.e., instability, volatility and irreproducibility) [2], but also has enabled the development of new methods, such as temperature-controlled IL dispersive liquid-liquid microextraction [3] or in-situ IL formation dispersive liquid-liquid microextraction [4,5].

A novel subclass of ILs named magnetic ionic liquids (MILs) has been recently presented as new extractant solvents for LPME techniques in few reports [6-13]. MILs include a paramagnetic cation or anion in their structure and possess similar physico-chemical properties to ILs but, additionally, show response under the application of an external magnetic field [14]. Specifically in LPME techniques, the reported studies have been performed with MILs

including [FeCl4-] [6-10], [FeCl3Br-] [8,11] or [MnCl4

2-] [12,13] anions. The

resulting paramagnetic behavior of these solvents enhances their manipulation providing remarkable advantages. For example, time-consuming centrifugation step is avoided in dispersive liquid-liquid microextraction [15] since phases separation is accomplished using a magnet [6-13], whereas in single-drop microextraction [16], the stability of the drop held by a small magnet is higher than the suspended drop in the needle tip of a syringe [8,13]. The main drawback regarding the use of MILs as extractant phase could be the related to the intrinsic hydrophilic nature and the hydrolytic instability of some halometallate-based MILs [14]. Thus, some investigations have addressed the


153

synthesis and characterization of new hydrophobic and stable MILs suitable for the treatment of aqueous samples [8,11-13]. In most of the reported works, a final dilution step of the enriched MIL drop [9,10,12,13] or a back-extraction procedure [11] are included after extraction in order to achieve a suitable medium for instrumental analysis by graphite furnace atomic absorption spectrometry [9] or liquid chromatography coupled to different detection systems (e.g., diode array, fluorimetric or UV-vis detectors) [10-13].

The purpose of this work is to present for the first time a headspace single-drop microextraction (HS-SDME) method based on the MIL 1-ethyl-3-methylimidazolium tetraisothiocyanatocobaltate (II) ([Emin]2[Co(NCS)4]). This MIL possesses hydrophilic properties but, in HS-SDME, the direct contact between the sample solution and the extractant phase is avoided and, therefore, also the limitations imposed by their miscibility. Nine chlorobenzenes were used as model analytes due to their volatility and good extractability in ILs [17-20]. After HS-SDME, the enriched MIL drop was directly analyzed using thermal desorption gas chromatography-mass spectrometry (TD-GC-MS). This system has been previously employed to analyze the IL 1-hexyl-3-methylimidazolium hexafluorophosphate ([C6mim][PF6]) after HS-SDME [20] but this is the first time that it is intended for the analysis of a MIL. The high boiling point of ILs and MILs provides advantages to perform stable and reproducible extractions in HS-SDME whereas this lack of volatility dirties the GC system and even blocks the column. The commercial TD system employed in this work enabled the desorption of model analytes from the MIL, while preventing the MIL from entering the GC system.

To the best of our knowledge, this is the first time that HS-SDME based on [Emin]2[Co(NCS)4] and coupled to TD-GC-MS has been reported. Various parameters of the proposed extraction method have been optimized by the use of a multivariate optimization strategy. Finally, the aforementioned method was evaluated in order to demonstrate its applicability to determine chlorobenzenes in real water samples.

Capítulo 2

154

2. Experimental part 2.1. Reagents and water samples

Target model analytes, namely 1,2-dichlorobenzene (1,2-DCB), 1,3-

dichlorobenzene (1,3-DCB), 1,4-dichlorobenzene (1,4-DCB), 1,2,3-trichlorobenzene (1,2,3-TCB), 1,2,4-trichlorobenzene (1,2,4-TCB), 1,3,5-trichlorobenzene (1,3,5-TCB), 1,2,3,4-tetrachlorobenzene (1,2,3,4-TeCB), 1,2,4,5-tetrachlorobenzene (1,2,4,5-TeCB), and pentachlorobenzene (PeCB), and 1,4-dibromobenzene (1,4-DBB) used as internal standard (IS) were all obtained from Riedel-de Haën (Seelze, Germany). Stock solutions of individual compounds (1000 mg L-1), and mix stock solutions containing the nine chlorobenzenes (10, 1 and 0.1 mg L-1) were prepared in LC grade acetonitrile from Sigma-Aldrich (St. Louis, MO, USA) and stored at 4 °C protected from light. Working solutions were daily prepared in ultrapure water (resistivity of 18.2 MΩ cm at 25 °C) from a Millipore Direct System Q5TM purification system from Ibérica S.A. (Madrid, Spain). Reactive grade NaCl was obtained from ACS Scharlau (Barcelona, Spain).

The MIL [Emim]2[Co(NCS)4] was supplied by the Department of Inorganic Chemistry of the University of Rostock (Germany). All details regarding the synthesis have been previously published [21].

Tap water was collected from the water-supplied network of the lab in the Department of Analytical Chemistry, Nutrition and Food Sciences of the University of Alicante (Spain). Pond water was collected in San Vicente del Raspeig (Alicante, Spain). Wastewater was from a wastewater treatment plant in A Coruña (Spain).

2.2. Materials and instrumentation

Rod shape NdFeB (Nd) magnets (N48, 3 mm diameter, 8 mm height) with a nickel-plated (Ni-Cu-Ni) coating from Supermagnete (Gottmadingen, Germany) were employed to support the drop of the MIL during HS-SDME. After extraction, Nd magnets were rinsed with distilled water and acetone, and dry at 240 ºC for 2 h before reused.


155

The sample compartment was a 22 mL glass vial with screw top (solid green Melamine cap and PTFE liner) from Supelco (Bellefonte, PA, USA). Sample solution was agitated during HS-SDME using a 1 cm length magnetic stir bar and a VWR® Advanced hot plate stirrer from VWR International (Radnor, PA, USA).

Thermogravimetric (TG) analysis was performed on a simultaneous TG-DTA system (model TGA/SDTA851e/SF/1100) from Mettler Toledo (Columbus, OH, USA). Measurements were done from 25 to 550 ºC at a heating rate of 10 ºC min-1 and under He atmosphere.

Magnetic susceptibility measurements were performed at 300 K in the magnetic field range -50 to 50 kOe using a MPMS XL (SQUID) magnetometer from Quantum Design (San Diego, CA, USA).

A Gerstel TDS 2 thermodesorption system equipped with a Gerstel CIS-4 cooled injection system programmable temperature vaporization inlet from Gerstel (Mülheim an der Ruhr, Germany) was used to carry out the TD process. A Gerstel TD glass tube (187 mm length, 4 mm I.D., 6 mm O.D.) and pesticide grade glass wool from Supelco were used to construct the TD device. This device was installed in a gas chromatograph (model 6890N) coupled to a mass spectrometer (model 5973) both from Agilent Technologies (Santa Clara, CA, USA). A capillary column HP-5ms (5% diphenyl-95% dimethylpolysiloxane, 30 m × 0.25 mm I.D., 0.25 µm film thickness) from Agilent Technologies (J&W GC columns) was employed for separation.

2.3. Headspace single-drop microextraction For HS-SDME, 20 mL of aqueous sample solution prepared with 15%

NaCl (w/v) were placed in a glass vial. 1 µL of MIL was held on the bottom of a rode shape Nd magnet (the lower magnet) and fixed to the vial cap using another Nd magnet (the upper magnet). Then, the MIL was exposed to the headspace of the aqueous sample solution stirred at 1500 rpm for 10 min at room temperature. After extraction, the lower magnet containing the enriched MIL was carefully dropped into a TD glass tube by removing the upper magnet and TD-GC-MS analysis was carried out. Figure 1 shows a scheme of the overall procedure.

Capítulo 2

156

Figure 1. Scheme of the overall proposed method.

2.4. TD-GC-MS conditions

The TD system was adjusted according to previously published works [20,22] as follows: splitless mode; desorption temperature, from 25 ºC (0.5 min) to 240 ºC at 60 ºC min-1; desorption time, 5 min; helium flow rate, 100 mL min-1; and transfer line temperature, 300 ºC. The desorbed compounds were cryofocused in the cooled injection system at 0 ºC. Then, the temperature was rapidly increased at 12 ºC s-1 up to 250 ºC, transferring the compounds to the GC column by operating in solvent vent mode for 1 min. The oven temperature was initially set at 40 ºC for 2 min, followed by a ramp of 5 ºC min-1 up to 130 ºC, and then at 10˚C min-1 to 240 ºC (5 min). Helium (99.999%) was used as the carrier gas at a constant flow rate of 1 mL min-1. The MS detector voltage was set at 1700 V. Electron impact ionization was used with ionization energy of 70 eV. The mass source and quadrupole were set at 230 and 150 ºC, respectively. Measurements were taken with a solvent delay of 2 min and in selected ion monitoring (SIM) mode at the following m/z ratios: 146 from minute 2 to 12 for dichlorobenzenes determination, 180 from minute 12 to 18 for trichlorobenzenes and IS, 216 from minute 18 to 22 for tetrachlorobenzenes and 250 from minute 22 to the end for PeCB. Figure 2

TD system

GC-MS

Upper magnet Lower magnet MIL

drop

TD tube

Glass wool

Lower magnet

MS


157

shows typical chromatograms after HS-SDME of a blank and a standard solution spiked at 0.7 µg L-1 with chlorobenzenes.

Figure 2. Total ion chromatograms obtained in the SIM mode after HS-SDME of a blank (green line) and a standard solution spiked at 0.7 µg L-1 with target analytes (black line). The concentration of IS in both chromatograms was 0.5 µg L-1.

1,3-DCB 1,4-DCB

1,2-DCB

1,3,5-TCB 1,2,4-TCB

1,2,3-TCB

1,2,4,5-TeCB

1,2,3,4-TeCB

PeCB

IS

Time (min)

Abu

ndan

ce (a

u)

Capítulo 2

158


A multivariate optimization strategy was carried out to determine the optimum conditions for HS-SDME. The statistical software NEMRODW® (“New Efficient Methodology for Research using Optimal Design”) from LPRAI (Marseille, France) was used to build the experimental design matrix and evaluate the results. As response functions, the peak area of individual compounds was considered.

3. Results and discussion 3.1. MIL characterization

[Emin]2[Co(NCS)4] has been deeply characterized in previous publications

[21,23]. Briefly, this MIL possesses good stability towards water and oxygen, good solubility in many solvents (e.g., water, acetonitrile, etc.) and low viscosity, even though doubly charged anion is present [21,23]. Figure S1 shows the magnetic susceptibility of the studied MIL at 300 K in the magnetic field range -50 to 50 kOe. The straight line obtained between the magnetic susceptibility and the applied magnetic field indicates the paramagnetic behavior of [Emin]2[Co(NCS)4] [24]. In addition, magnetic susceptibility was estimated from the slope of the linear fit to the data being 2.67 emu K mol-1, which it is similar to data previously reported in the literature for MILs containing the same type of anion [25].

Since a TD system was employed to analyze the MIL drop after HS-SDME, TG analysis was carried out in order to ensure that [Emin]2[Co(NCS)4]

was not evaporated under employed temperature conditions and, therefore, did not enter the GC column. According to TG curve (Figure S2), this MIL was very stable at temperatures lower than 250 ºC, with a minimum loss of mass (i.e., <5%) even at 300 ºC. Thus, it was concluded that [Emin]2[Co(NCS)4] was suitable for the proposed analytical method.


159

Figure S1. Magnetization of [Emin]2[Co(NCS)4] as a function of applied magnetic field at 300 K..

Figure S2. Thermogravimetric curve of [Emin]2[Co(NCS)4].

-500-400-300-200-100

0100200300400500

-60000 -40000 -20000 0 20000 40000 60000

M (e

mu

mol

-1)

H (Oe)

0

20

40

60

80

100

120

0 100 200 300 400 500 600

% m

ass

Temperature (ºC)

Capítulo 2

160

3.2. HS-SDME multivariate optimization A multivariate optimization strategy was carried out to optimize the

proposed HS-SDME method. In a first step, when a large number of factors could potentially affect extraction yield, the Plackett-Burman design was considered for screening purposes. The Plackett-Burman design is a fractional factorial design employed to study up to k=N-1 factors in N runs, where N is a multiple of 4 [26]. This design assumes that interaction between factors can be ignored so the main effects can be calculated with a reduced number of experiments [26].

On the basis of the literature and of the previous experience of the research group [18], the considered factors selected at two levels were: sample volume, MIL volume, extraction time, stirring speed and ionic strength (NaCl concentration). Extraction temperature is another key factor in some HS-SDME applications [27]. However, this factor was omitted in the present study since it showed non-significant effect for the extraction of chorobenzenes in our previous work [18]. Table S1 shows the experimental factors and levels considered in the Plackett-Burman design whereas Table S2 shows the matrix of experiments. The eight experiments were randomly performed using aqueous standards containing 1 µg L-1 of target analytes. The peak area of each chlorobenzene was employed as response function in order to evaluate the effect of different factors over each individual compound.

Table S1. Experimental factors and levels of the Plackettt-Burman design.

Factors Level

Low (-1) High (+1)

Sample volume (mL) 10 20

MIL volume (µL) 1 2


Stirring speed (rpm) 500 1500

Ionic strength, NaCl (%, w/v) 0 15


161

Table S2. Matrix of experiments of the Plackett-Burman design.

Run Sample

volume (mL) MIL volume

(µL) Extraction time (min)

Stirring speed (rpm)

Ionic strength, NaCl (%, w/v)

1 20 2 30 500 15

2 10 2 30 1500 0

3 10 1 30 1500 15

4 20 1 10 1500 15

5 10 2 10 500 15

6 20 1 30 500 0

7 20 2 10 1500 0

8 10 1 10 500 0 The data obtained were analyzed by ANOVA and the results were

visualized with the Pareto charts shown in Figure 3. The length of each bar was proportional to the influence of the corresponding factor, and negative and positive signals revealed whether the system responses decreased or increased, respectively, when passing from the lowest to the highest level of the corresponding factor. As shown in Figure 3, all analytes under study showed the same behaviour upon extraction. A reference vertical line would indicate the significance of a factor with 95% of probability. However, the vertical line did not appear in any of the charts, meaning that effects were far from the significance level. Although none of studied factors had a significant effect on the system responses, negative and positive signals indicated the most convenient level at which each factor should be fixed. Sample volume, stirring speed and ionic strength showed positive effects. On the one hand, greater sample volume involved greater amount of analytes, and therefore the responses increased. In addition, headspace volume was reduced when increasing sample volume (22 mL glass vials were employed in all experiments) and, according to theoretical calculations, low headspace volumes maximize extraction efficiency [27]. On the other hand, high stirring speed and NaCl salt content promoted the transfer of analytes from the sample solution to

Capítulo 2

162

the headspace due to a better diffusion and salting out effect [27]. MIL volume showed a negative effect. This fact revealed that, at working concentration, the amount of analytes extracted did not increase with extractant volume and a dilution effect occurred when passing from 1 to 2 µL of MIL. Finally, the effect of extraction time was also negative revealing a rapid and effective mass transfer in the proposed MIL-based HS-SDME system.

Figure 3. Pareto charts of the Plackett-Burman design.

223568.5

3193135.5

-1401911.25

-566356.75

519974

-4000000 -2000000 0 2000000 4000000

Ionic strength, NaCl (% w/v)



MIL volume (µL)

Sample volume (mL)

2888829.25

3912100

-1743573.25

-941480.25

659483.25

-5000000 -2500000 0 2500000 5000000




MIL volume (µL)

Sample volume (mL)

1,3-DCB

1,4-DCB


163

Figure 3 (cont.). Pareto charts of the Plackett-Burman design.

4936886.13

5512706.13

-2220755.13

-2541546.38

566730.38

-6000000 -3000000 0 3000000 6000000




MIL volume (µL)

Sample volume (mL)

4772737.63

5473674.63

-2302134.88

-2334711.38

1084707.63

-6000000 -3000000 0 3000000 6000000




MIL volume (µL)

Sample volume (mL)

8787159.63

8746084.88

-3262787.65

-5894292.38

1503128.88

-9000000 -4500000 0 4500000 9000000




MIL volume (µL)

Sample volume (mL)

1,3,5-TCB

1,2-DCB

1,2,4-TCB

Capítulo 2

164


9992260

9018790.25

-2526370.75

-7311119.5

803206.75

-10000000 -5000000 0 5000000 10000000




MIL volume (µL)

Sample volume (mL)

13438344.5

13324917.5

-4589740

-11216924

3332649.5

-15000000 -7500000 0 7500000 15000000




MIL volume (µL)

Sample volume (mL)

17173861.75

16639981.25

-5322536

-14753921.25 4042265.5

-20000000 -10000000 0 10000000 20000000




MIL volume (µL)

Sample volume (mL)

1,2,4,5-TeCB

1,2,3-TCB

1,2,3,4-TeCB


165


According to the results obtained from the Plackett-Burman design, the optimum conditions were finally established as follows: sample volume, 20 mL; MIL volume, 1 µL; extraction time, 10 min; stirring speed, 1500 rpm; and ionic strength, 15% NaCl (w/v).

3.3. Analytical figures of merit Calibration curve for each chlorobenzene was obtained by plotting the

ratio of the peak area of the analyte to peak area of the IS versus concentration of the analyte under the optimum experimental condition. The concentration of IS (1,4-DBB) was maintained constant at 0.5 µg L-1 in all experiments. The main analytical parameters of the proposed method are summarized in Table 1.

21995749.25

20859029.5

-8325283

-19585666.25

6721071.5

-25000000 -12500000 0 12500000 25000000




MIL volume (µL)

Sample volume (mL)

PeCB

Capítulo 2

166

Table 1. Main analytical parameters of the proposed method (MIL-based HS-SDME-TD-GC-MS).

a Correlation coefficient: number of calibration points in parentheses. b Coefficient of variation: mean value for three replicate analyses of aqueous standards spiked at indicated levels. c Limit of detection: calculated using blank signal plus three times its standard deviation. d Limit of quantification: calculated using blank signal plus ten times its standard deviation.

Analyte Working

range (µg L-1)

Slope (au µg-1 L) ra

CV (%)b LODc (ng L-1)

LOQd (ng L-1) 0.7 µg L-1 3 µg L-1

1,3-DCB 0.05-5 21.0±0.4 0.999 (6) 14 7 6.0 19.9 1,4-DCB 0.05-5 24.1±0.5 0.999 (6) 18 10 3.7 12.1 1,2-DCB 0.05-5 21.4±0.5 0.999 (6) 13 3 6.1 20.0

1,3,5-TCB 0.05-5 21.8±1.5 0.990 (6) 6 10 6.3 20.9 1,2,4-TCB 0.05-5 26.7±1.3 0.996 (6) 9 10 5.6 18.5 1,2,3-TCB 0.05-5 22.0±1.7 0.988 (6) 18 13 4.5 14.7

1,2,4,5-TeCB 0.05-5 34.2±1.8 0.995 (6) 4 6 7.6 25.1 1,2,3,4-TeCB 0.05-5 40.7±1.6 0.997 (6) 8 4 3.9 13.0

PeCB 0.05-5 60±3 0.996 (6) 8 5 6.3 20.8


167

The working range (from 0.05 to 5 µg L-1) showed good linearity with correlation coefficients ranging from 0.988 to 0.999 (Table 1). Student’s t-test was also applied to assess linearity [28], obtaining t calculated values from 12.8 (r=0.988; N=6) for 1,2,3-TCB, to 50.0 (r=0.999; N=6) for 1,3-DCB, being in all cases higher than the t tabulated value (t0.05,4=2.78). Hence, the null hypothesis of non-linear correlation for a 5% significance level and 4 degrees of freedom was rejected. The sensitivity of the instrumental measurements estimated by the slope of the calibration curves ranged between (21.0±0.4) au µg-1 L for 1,3-DCB and (60±3) au µg-1 L for PeCB. As can be seen from Table 1, the method provided the highest sensitivity for high molecular weight compounds (i.e., 1,2,4,5-TeCB, 1,2,3,4-TeCB and PeCB), which possess highest volatilities according to Henry’s law constants [29]. The repeatability of the measurements was evaluated by three replicate analysis of aqueous standards spiked at 0.7 and 3 µg L-1. The obtained coefficients of variation (CV) varied between 3 and 18% (Table 1).

Limit of detection (LOD) and limit of quantification (LOQ) (Table 1) were estimated according to 3Sb and 10Sb criteria [28], respectively, where Sb is the standard deviation of the blank. LODs ranged from 3.7 to 7.6 ng L-1, whereas LOQs varied from 12.1 to 25.1 ng L-1.

For comparative purposes, the characteristics of previously reported methods including IL-based HS-SDME for chlorobenzenes determination in water samples are summarized in Table 2.

Capítulo 2

168

Table 2. Comparison of the developed MIL-based HS-SDME-TD-GC-MS method with previously reported methods to determine chlorobenzenes in water samples.

MW, microwave-assisted; [Hmin][PF6], 1-hexyl-3-methylimidazolium hexafluorophosphate; LC-UV, liquid chromatography-UV detection; [Bmin][PF6], 1-butyl-3-methylimidazolium hexafluorophosphate; [Omin][PF6], 1-octyl-3-methylimidazolium hexafluorophosphate; GC-FID, gas chromatography-flame ionization detection; RT, room temperature.

Extraction method

Extractant solvent (volume)

Temp. (˚C)

Time (min)

Separation-detection

LOD (ng L-1)

Ref. 1,3- DCB

1,4- DCB

1,2- DCB

1,3,5- TCB

1,2,4- TCB

1,2,3- TCB

1,2,4,5- TeCB

1,2,3,4- TeCB PeCB

MW-HS-SDME

[Hmin][PF6] (5 µL)

- 20 LC-UV 32 39 24 22 22 16 16 16 - [17]

HS-SDME [Bmin][PF6] (5 µL)

25 37 LC-UV 152 203 102 122 122 102 102 102 - [18]

HS-SDME [Omin][PF6] (1 µL)

40 20 GC-FID - 500 100 - 500 - - - - [19]

HS-SDME [Hmin][PF6] (5 µL)

RT 37 TD-GC-MS 3 2 2 4 2 2 2 2 1 [20]

HS-SDME [Emin]2[Co(NCS)4] (1 µL)

RT 10 TD-GC-MS 6.0 3.7 6.1 6.3 5.6 4.5 7.6 3.9 6.3 This work


169

As can be seen, the proposed method showed comparable or lower LODs than previous works and used shorter extraction times. In addition, extractions were performed at room temperature, thus saving energy. Finally, MIL-based HS-SDME showed unique advantages related with the easy manipulation of the solvent and the low volume employed (i.e., 1 µL), which reduced the generation of wastes.

3.4. Analysis of water samples

The applicability of the proposed method to determine chlorobenzenes in real water samples was evaluated. Three water samples (namely, tap water, pond water and wastewater) were employed to assess matrix effects using recovery studies. It should be noted that in previous analyses with the proposed method, none of the selected water samples had initial detectable concentrations of target analytes. Three replicated analyses of each water sample were carried out at 0.7 and 3 µg L-1 spiking levels. Table 3 shows the relative recoveries determined as the ratio of the signals found after HS-SDME in real and ultrapure water samples spiked at the same concentration levels. Results showed relative recoveries varying from 88 and 110% for tap water, between 82 and 113% for pond water, and between 84 and 114% for wastewater. CV values ranged between 1 and 23%. According to these results, we can conclude that matrix effects were not significant for the determination of chlorobenzenes in the three selected water samples.

Capítulo 2

170

Table 3. Relative recoveries and CV values (in parentheses) obtained for the target analytes in the three studied real water samples.

Analyte Relative recoveries (%) and CV values in parentheses (%)a

Tap water Pond water Wastewater

0.7 µg L-1 3 µg L-1 0.7 µg L-1 3 µg L-1 0.7 µg L-1 3 µg L-1

1,3-DCB 92 (8) 94 (15) 84 (17) 95 (18) 103 (3) 89 (18) 1,4-DCB 99 (6) 94 (12) 91 (13) 105 (19) 114 (5) 90 (13) 1,2-DCB 92 (11) 89 (14) 87 (13) 100 (10) 101 (8) 96 (20)

1,3,5-TCB 100 (20) 94 (13) 84 (18) 106 (12) 93 (5) 87 (15) 1,2,4-TCB 94 (16) 88 (13) 85 (13) 111 (23) 97 (12) 98 (9) 1,2,3-TCB 97 (18) 95 (10) 91 (12) 113 (16) 113 (20) 98 (12)

1,2,4,5-TeCB 96 (15) 94 (12) 89 (18) 108 (23) 86 (18) 96 (5) 1,2,3,4-TeCB 101 (13) 95 (8) 92 (13) 99 (22) 84 (17) 101 (8)

PeCB 110 (5) 104 (1) 82 (15) 90 (19) 92 (17) 98 (22)

a Three replicate analysis at indicated levels.

4. Conclusions

A new sensitive and easy to handle MIL-based HS-SDME method using [Emin]2[Co(NCS)4] as extractant solvent has been proposed in this work. [Emin]2[Co(NCS)4] is a water-miscible solvent but it has been successfully applied for the extraction of chlorobenzenes from water samples due to the direct contact between sample solution and extractant phase is avoided in HS-SDME. The paramagnetic properties of [Emin]2[Co(NCS)4] provided high stability to the MIL drop held by a Nd magnet and facilitated the manipulation. On the other hand, the commercial TD system employed in this work allowed the easy and direct determination of analytes by GC-MS, without needing further dilution or back-extraction steps, opening new possibilities and broaden future applications for MIL-based LPME techniques.

The proposed method has been applied for the extraction of chlorobenzenes from water samples as model analytical application. Results


171

showed that comparable or even lower LOD values than previously reported works based on analogous systems (i.e., IL-based HS-SDME) were obtained with shorter extraction times. Finally, our study demonstrates the ability of the proposed method to determine chlorobenzenes at trace levels in real water samples.

Acknowledgements

The authors would like to thank Generalitat Valenciana (projects n. GVA/2014/096 and PROMETEO/2013/038) and Ministerio de Economía, Industria y Competitividad (project n. CTQ2016-79991-R) for the financial support. The authors would also like to thank Dr. Martin Köckerling from the Department of Inorganic Chemistry of the University of Rostock (Germany) for the MIL supply. E. Fernández thanks Ministerio de Educación for her FPU grant (FPU13/03125).

References

[1] P. Sun, D.W. Armstrong, Ionic liquids in analytical chemistry, Anal. Chim. Acta 661 (2010) 1-16.

[2] E. Aguilera-Herrador, R. Lucena, S. Cárdenas, M. Valcárcel, The roles of ionic liquids in sorptive microextraction techniques, Trends Anal. Chem. 29 (2010) 602-616.

[3] Q. Zhou, H. Bai, G. Xie, J. Xiao, Temperature-controlled ionic liquid dispersive liquid phase micro-extraction, J. Chromatogr. A 1177 (2008) 43-49.

[4] M. Baghdadi, F. Shemirani, In situ solvent formation microextraction based on ionic liquids: A novel sample preparation technique for determination of inorganic species in saline solutions, Anal. Chim. Acta 634 (2009) 186-191.

[5] C. Yao, J.L. Anderson, Dispersive liquid-liquid microextraction using an in situ metathesis reaction to form an ionic liquid extraction phase for the preconcentration of aromatic compounds from water, Anal. Bioanal. Chem. 395 (2009) 1491-1502.

Capítulo 2

172

[6] Y. Wang, Y. Sun, B. Xu, X. Li, R. Jin, H. Zhang, D. Song, Magnetic ionic liquid-based dispersive liquid-liquid microextraction for the determination of triazine herbicides in vegetable oils by liquid chromatography, J. Chromatogr. A 1373 (2014) 9-16.

[7] Y. Wang, Y. Sun, B. Xu, X. Li, X. Wang, H. Zhang, D. Song, Matrix solid-phase dispersion coupled with magnetic ionic liquid dispersive liquid-liquid microextraction for the determination of triazine herbicides in oilseeds, Anal. Chim. Acta 888 (2015) 67-74.

[8] K.D. Clark, O. Nacham, H. Yu, T. Li, M.M. Yamsek, D.R. Ronning, J.L. Anderson, Extraction of DNA by magnetic ionic liquids: Tunable solvents for rapid and selective DNA analysis, Anal. Chem. 87 (2015) 1552-1559.

[9] X. Wang, G. Xu, P. Chen, X. Liu, Y. Fang, S. Yang, G. Wang, Arsenic speciation analysis in environmental water, sediment and soil samples by magnetic ionic liquid-based air-assisted liquid-liquid microextraction, RSC Adv. 6 (2016) 110247-110254.

[10] T. Chatzimitakos, C. Binellas, K. Maidatsi, C. Stalikas, Magnetic ionic liquid in stirring-assisted drop-breakup microextraction: Proof-of-concept extraction of phenolic endocrine disrupters and acidic pharmaceuticals, Anal. Chim. Acta 910 (2016) 53-59.

[11] M.J. Trujillo-Rodríguez, O. Nacham, K.D. Clark, V. Pino, J.L. Anderson, J.H. Ayala, A.M. Afonso, Magnetic ionic liquids as non-conventional extraction solvents for the determination of polycyclic aromatic hydrocarbons, Anal. Chim. Acta 934 (2016) 106-113.

[12] H. Yu, J. Merib, J.L. Anderson, Faster dispersive liquid-liquid microextraction methods using magnetic ionic liquids as solvents, J. Chromatogr. A 1463 (2016) 11-19.

[13] J. An, K.L. Rahn, J.L. Anderson, Headspace single drop microextraction versus dispersive liquid-liquid microextraction using magnetic ionic liquid extraction solvents, Talanta 167 (2017) 268-278.

[14] K.D. Clark, O. Nacham, J.A. Purslow, S.A. Pierson, J.L. Anderson, Magnetic ionic liquids in analytical chemistry: A review, Anal. Chim. Acta 934 (2016) 9-21.


173

[15] M. Rezaee, Y. Assadi, M.-R. Milani Hosseini, E. Aghaee, F. Ahmadi, S. Berijani, Determination of organic compounds in water using dispersive liquid-liquid microextraction, J. Chromatogr. A 1116 (2006) 1-9.

[16] M.A. Jeannot, F.F. Cantwell, Mass transfer characteristics of solvent extraction into a single drop at the tip of a syringe needle, Anal. Chem. 69 (1997) 235-239.

[17] L. Vidal, C.E. Domini, N. Grané, E. Psillakis, A. Canals, Microwave-assisted headspace single-drop microextration of chlorobenzenes from water samples, Anal. Chim. Acta 592 (2007) 9-15.

[18] L. Vidal, E. Psillakis, C.E. Domini, N. Grané, F. Marken, A. Canals, An ionic liquid as a solvent for headspace single drop microextraction of chlorobenzenes from water samples, Anal. Chim. Acta 584 (2007) 189-195.

[19] F. Zhao, S. Lu, W. Du, B. Zeng, Ionic liquid-based headspace single-drop microextraction coupled to gas chromatography for the determination of chlorobenzene derivatives, Microchim. Acta 165 (2009) 29-33.

[20] A. Chisvert, I.P. Román, L. Vidal, A. Canals, Simple and commercial readily-available approach for the direct use of ionic liquid-based single-drop microextraction prior to gas chromatography. Determination of chlorobenzenes in real water samples as model analytical application, J. Chromatogr. A 1216 (2009) 1290-1295.

[21] T. Peppel, M. Köckerling, M. Geppert-Rybczyńska, R. V Ralys, J.K. Lehmann, S.P. Verevkin, A. Heintz, Low-viscosity paramagnetic ionic liquids with doubly charged [Co(NCS)4]2- ions, Angew. Chemie 49 (2010) 7116-7119.

[22] L. Vidal, M. Ahmadi, E. Fernández, T. Madrakian, A. Canals, Magnetic headspace adsorptive extraction of chlorobenzenes prior to thermal desorption gas chromatography-mass spectrometry, Anal. Chim. Acta 971 (2017) 40-47.

[23] M. Geppert-rybczyn, J.K. Lehmann, T. Peppel, M. Ko, A. Heintz, Studies of physicochemical and thermodynamic properties of the paramagnetic 1-alkyl-3-methylimidazolium ionic liquids

Capítulo 2

174

(EMIm)2[Co(NCS)4] and (BMIm)2[Co(NCS)4], J. Chem. Eng. Data 55 (2010) 5534-5538.

[24] E. Santos, J. Albo, A. Rosatella, C.A.M. Afonso, A. Irabien, Synthesis and characterization of Magnetic Ionic Liquids (MILs) for CO2 separation, J. Chem. Technol. Biotechnol. 89 (2014) 866-871.

[25] R.E. Del Sesto, T.M. McCleskey, A.K. Burrell, G.A. Baker, J.D. Thompson, B.L. Scott, J.S. Wilkes, P. Williams, Structure and magnetic behavior of transition metal based ionic liquids, Chem. Commun. 8 (2008) 447-449.

[26] D.C. Montgomery, Design and Analysis of Experiments, 7th ed., Wiley, New Jersey (USA), 2009.

[27] M.A. Jeannot, A. Przyjazny, J.M. Kokosa, Single drop microextraction-Development, applications and future trends, J. Chromatogr. A 1217 (2010) 2326-2336.

[28] J.N. Miller, J.C. Miller, Statistics and Chemometrics for Analytical Chemistry, 5th ed., Pearson Prentice Hall, London (UK), 2005.

[29] D. Mackay, W.Y. Shiu, A critical review of Henry´s law constants for chemicals of environmental interest, J. Phys. Chem. Ref. Data 10 (1981) 1175-1189.

Capítulo 3


177

Complexation-mediated electromembrane extraction of highly polar basic drugs - a fundamental study with catecholamines in

urine as model system

Elena Fernándeza, Linda Vårdalb, Lorena Vidala, Antonio Canalsa, Astrid Gjelstadb, Stig Pedersen-Bjergaardb

a Department of Analytical Chemistry, Nutrition and Food Sciences and University Institute of Materials, University of Alicante, P.O. Box 99, E-03080 Alicante, Spain. b School of Pharmacy, University of Oslo, P.O. Box 1068 Blindern, 0316 Oslo, Norway.

Received: 23 January 2017 / Revised: 22 March 2017 / Accepted: 19 April 2017 / Published online: 28 April 2017

Abstract

Complexation-mediated electromembrane extraction (EME) of highly polar basic drugs (log P<-1) was investigated for the first time with the catecholamines epinephrine, norepinephrine, and dopamine as model analytes. The model analytes were extracted as cationic species from urine samples (pH 4), through a supported liquid membrane (SLM) comprising 25 mM 4-(trifluoromethyl)phenylboronic acid (TFPBA) in bis(2-ethylhexyl) phosphite (DEHPi), and into 20 mM formic acid as acceptor solution. EME was performed for 15 minutes, and 50 V was used as extraction voltage across the SLM. TFPBA served as complexation reagent, and selectively formed boronate esters by reversible covalent binding with the model analytes at the sample/SLM interface. This enhanced the mass transfer of the highly polar model analytes across the SLM, and EME of basic drugs with log P in the

Anal Bioanal Chem (2017) DOI: 10.1007/s00216-017-0370-2

Capítulo 3

178

range -1 to -2 was shown for the first time. Meanwhile most matrix components in urine were unable to pass the SLM. Thus, the proposed concept provided highly efficient sample clean-up and the system current across the SLM was kept below 50 μA. Finally, the complexation-mediated EME concept was combined with ultra-high performance liquid chromatography coupled to tandem mass spectrometry and evaluated for quantification of epinephrine and dopamine. Standard addition calibration was applied to a pooled human urine sample. Calibration curves using standards between 25 and 125 µg L-1 gave a high level of linearity with a correlation coefficient of 0.990 for epinephrine and 0.996 for dopamine (N=5). The limit of detection, calculated as three times signal-to-noise ratio, was 5.0 µg L-1 for epinephrine and 1.8 µg L-1 for dopamine. The repeatability of the method, expressed as coefficient of variation, was 13% (n=5). The proposed method was finally applied for the analysis of spiked pooled human urine sample, obtaining relative recoveries of 91 and 117% for epinephrine and dopamine, respectively.

Keywords: electromembrane extraction, polar analytes, urine samples, catecholamines

1. Introduction

Electromembrane extraction (EME) is a miniaturized extraction technique evolved from hollow fiber liquid-phase microextraction (HF-LPME) [1]. In EME, charged analytes are extracted from aqueous sample, through an organic solvent immobilized as a supported liquid membrane (SLM) in the pores of a polymeric hollow fiber, and into an acceptor solution located in the lumen of the fiber [2]. An electrical potential difference is employed as driving force for the electrokinetic migration of analytes across the SLM. A power supply provides a DC potential between two electrodes placed in the sample and acceptor solution, respectively. For the extraction of basic analytes, the anode (positively charged electrode) is placed into sample whereas the cathode (negatively charged electrode) is placed into the acceptor solution. For the extraction of acidic analytes, the direction of the electrical field is reversed, the cathode is located in the sample and the anode is located in the acceptor


179

solution. The pH of both sample and acceptor solution has to be controlled to ensure full ionization of the target analytes. Major advantages of EME include [3,4]: low consumption of organic solvents; shorter extraction times than HF-LPME due to the enhancement of mass transport by the force of the electrical potential; efficient sample clean-up and feasibility of direct extraction from untreated complex matrices; easy extraction selectivity modulation by changes in the magnitude and direction of the electrical potential; high preconcentration capacity; direct compatibility with a wide range of analytical instruments; simple and low cost equipment; and possibilities of downscaled format (i.e., microchip devices) and automation.

Experimental parameters such as the SLM composition, extraction voltage, extraction time, pH of sample and acceptor solutions, salt effect, and sample stirring speed strongly affect EME performance, and are normally optimized in different applications [2,3]. The selection of appropriate solvent within the pores of the fiber is a critical task of the technique. Some important properties of the solvent to consider are immiscibility with water to prevent losses by dissolution, low volatility to avoid evaporation during extraction, low viscosity to ensure high diffusion coefficients across SLM, good extractability and high partition coefficient of the target analytes, and certain dipole moment or conductivity to support current flow in the system [3,5]. For the EME of nonpolar (log P>2) basic drugs, 2-nitrophenyl octylether (NPOE) [6-14] has been the most employed solvent, although 1-ethyl-2-nitrobenzene (ENB) [15-18] and 1-isopropyl-4-nitrobenzene (IPNB) [19,20] have been alternatively proposed, performing extractions at low voltages. NPOE, ENB and IPNB possess low water solubility, high boiling point, and are able to form dipole-dipole and hydrogen bonding interactions with positively charged analytes, thus being suitable solvents to create efficient SLMs [20,21]. The extraction of polar (log P<2) basic drugs is more challenging since these species are less prone to cross the hydrophobic SLM under the influence of an electrical field. In this case, the presence of carriers in the SLM is compulsory to promote the analyte transfer and to increase EME efficiency. Among tested carriers, di(2-ethylhexyl) phosphate (DEHP) and tri(2-ethylhexyl) phosphate (TEHP) have been the most popular ones [21]. DEHP forms ion-pairs with positively charged basic drugs, whereas TEHP is a non-ionic carrier interacting with

Capítulo 3

180

charged analytes mainly by dipole-dipole and hydrogen interactions. DEHP has been more efficient than TEHP for the extraction of the most polar basic drugs (0.01<log P<1.8) [21]. However, DEHP suffers from some drawbacks related to the increase of the electrical conductance of the SLM and extraction of back-ground electrolyte ions and other ionic sample components, leading to high system currents [22]. Very recently, a new SLM based on bis(2-ethylhexyl) phosphite (DEHPi) has been discovered as a good candidate for the extraction of polar (log P values between -0.40 and 1.32) basic analytes from plasma samples [22]. DEHPi was compared with SLMs based on DEHP and TEHP, and DEHPi provided lower currents and higher system stability [22].

Experiences with EME of basic drug substances of very high polarity (-1> log P>-2) have not yet been reported in the literature, and therefore a fundamental study on this was addressed in the present work. The catecholamines dopamine (DA) (log P=-0.99), epinephrine (E) (log P=-1.37), and norepinephrine (NE) (log P=-1.85) were selected as model analytes [23]. In order to enhance their mass transfer across the SLM, and to maintain an acceptable level of selectivity and sample clean-up from biological fluids, different analogues of phenylboronic acid (PBA) were added to the EME system as selective complexation reagents for the catecholamines. Operational parameters for this conceptually new type of complexation-mediated EME system were studied and optimized to obtain fundamental experience and knowledge. Special emphasis was devoted to recovery, current stability, and sample clean-up. The optimized EME system was finally combined with ultra-high performance liquid chromatography coupled to tandem mass spectrometry (UHPLC-MS/MS), and evaluated for the quantification of DA and E in human urine.


2.1. Chemicals

Dopamine hydrochloride, epinephrine hydrochloride, norepinephrine bitartrate, 1,4-benzodioxane-6-boronic acid, 4-(trifluoromethyl)phenylboronic acid (TFPBA), m-tolylboronic acid, 4-(benzyloxy)phenylboronic acid, 4-


181

(dimethylamino)phenylboronic acid, 4-(trans-2-carboxyvinyl)phenylboronic acid, DEHP, DEHPi, formic acid, and sodium 1-heptanesulfonate were all purchased from Sigma-Aldrich (St. Louis, MO, USA). PBA and NPOE were obtained from Fluka (Buchs, Switzerland). Hydrochloric acid, phosphoric acid, sodium dihydrogen phosphate monohydrate, disodium hydrogen phosphate dodecahydrate, trisodium phosphate dodecahydrate, and methanol were supplied by Merck (Darmstadt, Germany). The ultrapure water (resistivity of 18.2 MΩ cm at 25 ºC) employed for preparing aqueous solutions was obtained with a Milli-Q water purification system (Molsheim, France). 2.2. Solutions and urine samples

Stock solutions of E, NE and DA were prepared at 1000 mg L-1 in methanol and stored at 5 °C protected from light. Aqueous working solutions were daily prepared by proper dilution of stock solutions with selected background electrolyte (i.e., 10 mM hydrochloric acid or 20 mM phosphate buffer). Solutions of 1 mg L-1 containing the three analytes were employed in initial experiments and EME optimization.

Urine samples were collected from healthy volunteers in sterilized containers and kept at 5 ºC before analysis. Urine samples were diluted with 20 mM phosphate buffer of predetermined pH (volume ratio 1:1) before EME experiments.


Two chromatographic systems were employed for EME optimization and method evaluation, respectively. For EME optimization, chromatographic analysis was performed by high performance liquid chromatography coupled to ultraviolet detection (HPLC-UV). The chromatographic system containing a degasser, a binary pump, and an autosampler (all of 1200 series) was from Agilent Technologies (Santa Clara, CA, USA). Gemini C18 column (150 mm x 2 mm I.D, 5 µm particle size) from Phenomenex (Torrance, CA, USA) was employed for separation. The injection volume was 10 µL. Analytes were

Capítulo 3

182

eluted in gradient mode using mobile phases A and B. Mobile phase A consisted of 95% water phase (20 mM formic acid and 5 mM sodium 1-heptanesulfonate in ultrapure water) and 5% methanol. Mobile phase B consisted of 95% methanol and 5% water phase (20 mM formic acid and 5 mM sodium 1-heptanesulfonate in ultrapure water). Elution program was as follows: mobile phase B was increased from 3 to 35% within 12 min. Then, mobile phase B was further increased to 80% in 0.5 min and this condition was kept for 3.5 min. Finally, the mobile phase composition was returned to the starting conditions and held constant for 4 min before next injection. The total analysis time was 20 min with a flow rate of 0.4 mL min-1. The UV detector was set at 280 nm.

Method evaluation was carried out using UHPLC-MS/MS. The chromatographic system comprised a Dionex UltiMate 3000 RS pump, autosampler, and column compartment followed by a LTQ XL linear ion trap mass spectrometer from Thermo Scientific (San Jose, CA, USA). Chromatographic separation was achieved on an Acquity UPLC® HSS T3 column (100 mm x 2.1 mm I.D, 1.8 µm particle size) from Waters (Wexford, Ireland) kept at 40 ºC. The injection volume was 5 µL. Mobile phase A contained 95% water phase (20 mM formic acid in ultrapure water) and 5% methanol. Mobile phase B contained 95% methanol and 5% water phase (20 mM formic acid in ultrapure water). The linear gradient elution was programmed from 1 to 80% of mobile phase B in 1.5 min. 80% of mobile phase B was kept for 1 min before changing back to the starting conditions for equilibration. The total analysis time was 5.5 min with a flow rate of 0.3 mL min-1. MS/MS detection was acquired in the selected reaction monitoring (SRM) mode with electrospray ionization in the positive mode. Transitions (m/z) 184166 and 154137 were monitored for E and DA, respectively, for quantitative purposes. NE was excluded from the method evaluation in this conceptual work (i.e., UHPLC-MS/MS) since its quantification in the concentration range of interest (i.e., µg L-1 level) was not achieved. The source fragmentation energy was 35 V and the collision energy was 15% for E and 17% for DA.


183

2.4. EME set-up and procedure

The sample compartment was a 2 mL glass vial with screw cap from Supelco (Bellefonte, PA, USA). The hollow fiber used as the support for the organic solvent and for housing the acceptor solution was a PP Q3/2 polypropylene hollow fiber from Membrana (Wuppertal, Germany) with an internal diameter of 1.2 mm, wall thickness of 200 µm, and pore size of 0.2 µm. A Thermomixer Comfort agitator from Eppendorf (Hamburg, Germany) was used to agitate the extraction unit during EME. Platinum wires with 0.5 mm of diameter were used as electrodes. The electric potential was generated by a DC power supply (model ES 0300-0.45) from Delta Electronika (Zierikzee, The Netherlands). Current was monitored during EME using an Agilent U1253B True Rms Oled multimeter.

EME was performed according to the following procedure: 1 mL of sample solution was placed into 2 mL glass vial. The polypropylene hollow fiber was cut in a 2.5-cm piece whose lower end was sealed by mechanical pressure. The upper end was connected by heat to a 2.2-cm length pipette tip (Finntip 200 Ext from Thermo Scientific) acting as guiding tube. The hollow fiber was dipped for 5 s in the organic solvent used as SLM and the excess of solvent was thereafter removed with a medical wipe. Via guiding tube, 25 µL of acceptor solution was filled into the lumen of the hollow fiber with a microsyringe. Subsequently, the hollow fiber was inserted through the vial cap and introduced in the sample. Finally, the cathode was placed in the acceptor solution and the anode in the sample. The electrodes were connected to the power supply and the extraction unit was agitated at 900 rpm for a predetermined time. After EME, acceptor solution was collected with a microsyringe for its final injection in the corresponding chromatographic system (i.e., HPLC-UV for optimization studies and UHPLC-MS/MS to evaluate the method).

Capítulo 3

184

2.5. Calculations

The EME recovery was calculated using the following equation:

Recovery (%)= CaVa

CsVs×100

where Ca is the final concentration of the analyte in the acceptor solution, Cs is the initial analyte concentration in the sample solution, Va is the volume of the acceptor solution, and Vs is the volume of the sample.


3.1. Experiments based on conventional EME

First, experiments were performed using pure NPOE as SLM. The catecholamines were dissolved in 10 mM hydrochloric acid (pH 2), and this solution served as sample. EME was operated at 300 V. After 5 min of extraction, no analytes were detected by HPLC-UV in the acceptor solution. The catecholamines were then dissolved in 20 mM phosphate buffer (pH 5), and with this solution serving as sample, EME was repeated under equal conditions. However, also in this case, no extraction of the catecholamines was observed. The inefficiency of NPOE was expected. NPOE is well-known to efficiently extract non-polar basic compounds (log P˃2) by strong dipole and hydrogen bonding interactions. On the other hand, the extraction of polar analytes with low affinity to the SLM generally requires the use of hydrophobic ion-pair reagents, such as DEHP, acting as carriers [21].

DEHP has been frequently combined with NPOE for the extraction of polar substances, since its ability to form complexes with positively charged species facilitates their transfer into the SLM [21]. A SLM based on NPOE with 10% (w:w) of DEHP was tested for the catecholamines using an extraction voltage of 25 V. Standard solutions of pH 2 and 5 (10 mM hydrochloric acid and 20 mM phosphate buffer, respectively) were subjected to EME for 5 min. Surprisingly, the analytes were not found in the corresponding acceptor solutions, even not at trace level. Thus, the SLM comprising a mixture


185

of DEHP and NPOE appeared to be insufficient for mass transfer of the highly polar catecholamines.

DEHPi has been recently demonstrated as SLM for extraction of polar basic drugs in the log P range from -0.40 to 1.32 [22]. DEHPi was also tested in the current work for catecholamines using an applied voltage of 50 V. After 5 min of extraction from an aqueous standard solution of pH 2 (10 mM hydrochloric acid), catecholamines were now detected in the acceptor solution. The extraction recoveries were 0.3% for E, 0.4% for NE, and 0.8% for DA. The experiment was repeated with 10 min extraction time, and extraction recoveries increase to 0.5% for E, 0.7% for NE, and 1.6% for DA. However, a more significant improvement was observed using a standard solution of pH 5 (20 mM phosphate buffer), and recoveries were now 3% for E, 6% for NE, and 14% for DA after 10 min of extraction. The pH dependence observed was unexpected since DEHPi is not able to form ionic interactions under normal pH conditions [22]. The enhancement in extraction performance at higher pH was hypothesized to be due to the presence of small amounts of ionic oxidation products in DEHPi. Thus, special attention should be paid in the manipulation of DEHPi, using closed containers to avoid its progressive oxidation as far as possible.

3.2. Experiments based on complexation-mediated EME

The molecular structures of the catecholamines include two phenolic groups in ortho position as a common feature. PBA and derivatives possess a high affinity to complex these phenols, forming boronate esters by reversible covalent binding (Figure 1). Based on this type of complexation, previous publications [24-26] have reported the ability of PBA derivatives to facilitate transport of diol containing species (e.g., DA, glucoside, fructose) through SLMs under passive diffusion conditions. This concept was transferred to EME in the present work, and tested under electrokinetic migration conditions. The idea was to enhance the mass transfer of catecholamines due to selective complexation, while suppressing the general mass transfer of cationic matrix components.

Capítulo 3

186

Figure 1. PBA complexation of diol groups.

In a first experiment, PBA was dissolved in standard solution of pH 5 at a

concentration of 5 mM. EME was performed for 10 min at 50 V using DEHPi as SLM. Under these conditions, recoveries were 4% for E, 10% for NE, and 20% for DA. The improvement in extraction efficiency, especially for NE and DA, was attributed to decreased polarity of these molecules via complexation (Figure 1). Based on this positive finding, the potential for complexation-mediated EME was studied in more detail below.

3.3. Optimization

3.3.1. Type of complexing agent

PBA and six different derivatives (namely: 1,4-benzodioxane-6-boronic acid; TFPBA; m-tolylboronic acid; 4-(benzyloxy)phenylboronic acid; 4-(dimethylamino) phenylboronic acid; and 4-(trans-2-carboxyvinyl)phenylboronic acid) were investigated using DEHPi as SLM. For stepwise development of experiences, optimization was performed with aqueous standard solutions. The complexing reagents were dissolved in the sample solution or in DEHPi depending on their polarity and water miscibility. Thus, PBA (log P=1.64) and 1,4-benzodioxane-6-boronic acid (log P=0.95) were added to the aqueous sample, and with these reagents complexation was expected in the bulk sample. In contrast, TFPBA (log P=2.52), m-tolylboronic acid (log P=2.11), and 4-(benzyloxy)-phenylboronic acid (log P=3.16) were dissolved in the SLM. With these reagents, complexation was expected at the sample/SLM interface. The use of equal amounts (moles) of the different reagents was considered necessary in order to compare their net effect on EME. Therefore, reagents in the aqueous standard (1 mL) were dissolved at a concentration of 1 mM, whereas reagents in the SLM (approximately 20 µL) were dissolved at a concentration of 50 mM. The dissolution of 4-

B

OH

OH

HO

HO

B

O

O

B

OH

OH

HO

HO

B

O

O

B

OH

OH

HO

HO

B

O

O

+ 2 H2O+


187

(dimethylamino)phenylboronic acid (log P=1.90) and 4-(trans-2-carboxyvinyl)phenylboronic acid (log P=1.99) in aqueous phase or DEHPi was not achieved at selected concentrations, and these derivatives were therefore discarded. The effect of the different complexing reagents on EME of catecholamines is shown in Figure 2.

Figure 2. Effect of complexing reagent. Extraction conditions: concentration of analytes, 1 mg L-1; sample pH, 5; SLM, DEHPi; applied voltage, 50 V; extraction time, 10 min; acceptor solution, 20 mM formic acid. Error bars represent the standard deviation of three replicated analysis.

As observed, higher recoveries were obtained with TFPBA dissolved in DEHPi, and therefore this reagent was selected for further investigations together with PBA. TFPBA and PBA were both tested with NPOE as SLM, but these EME systems were not efficient. Thus, DEHPi was used as SLM in all remaining experiments.

0

5

10

15

20

25

30

Epinephrine Norepinephrine Dopamine

Rec

over

y (%

)

Phenylboronic acid

1,4-benzodioxane-6-boronic acid

m-tolylboronic acid

4-(trifluoromethyl)phenylboronic acid

4-(benzyloxy)phenylboronic acid

Capítulo 3

188

3.3.2. Concentration of complexing agent

Different concentrations of PBA in the sample solution (i.e., 0, 0.5, 1, 3, 5, and 10 mM) and TFPBA in SLM (i.e., 0, 25, 50, and 250 mM) were evaluated. As shown in Figure 3a, the effect of PBA on EME performance was practically negligible at concentrations of 0.5 and 1 mM. In these experiments, the molar concentration of PBA was 30-60 times higher than the analyte concentrations used in the experiment (1 mg L-1). However, a significant increase in extraction was observed at 3, 5 and 10 mM, especially for NE and DA. System current measurements revealed that current increased with PBA concentration, although it was kept below 50 µA in all cases [22]. Finally, 3 mM PBA was selected as optimum value since extraction recoveries were comparable to those obtained at higher concentrations (i.e., 5 and 10 mM), but the EME system was more stable. Regarding TFPBA, Figure 3b shows an enhancement in extraction performance as the reagent concentration increased. However, as observed with PBA, the system current increased with the concentration of the complexing reagent exceeding 50 µA at 250 mM. Finally, 25 mM TFPBA was selected as a compromise value.


189

(a)

(b)

Figure 3. Effect of (a) PBA concentration, and (b) TFPBA concentration. Extraction conditions: concentration of analytes, 1 mg L-1; sample pH, 5; SLM, DEHPi; applied voltage, 50 V; extraction time, 10 min; acceptor solution, 20 mM formic acid. Error bars represent the standard deviation of three replicated analysis.

The EME of catecholamines using simultaneously 3 mM PBA dissolved in sample solution (i.e., aqueous standard) and 25 mM TFPBA dissolved in DEHPi was also tested. Recoveries were not significantly different to those

0

5

10

15

20

25


Rec

over

y (%

) 0 mM0.5 mM1 mM3 mM5 mM10 mM

05

1015202530354045


Rec

over

y (%

)

0 mM25 mM50 mM250 mM

Capítulo 3

190

obtained with complexing reagents separately and system current increased. Therefore, the simultaneous use of PBA and TFPBA was discarded.

Finally, selected optimum conditions (i.e., 3 mM PBA or 25 mM TFPBA dissolved in sample solution or DEHPi, respectively) were evaluated in a real urine sample diluted with 20 mM phosphate buffer pH 5 (volume ratio 1:1). Higher recoveries were obtained with 25 mM TFPBA (Figure S1). Additionally, a general increase in system current was observed when EME was conducted from the real samples compared to aqueous standards. However, the increase in current was lower and kept below 50 µA for TFPBA. According to these results, TFPBA was finally selected as complexing reagent for EME of the catecholamines.

Figure S1. Extraction performance in urine sample using DEHPi as SLM in the absence of complexing reagent, with 3 mM PBA dissolved in sample solution and with 25 mM TFPBA dissolved in the SLM, respectively. Extraction conditions: concentration of analytes, 1 mg L-1; applied voltage, 50 V; extraction time, 10 min; acceptor solution, 20 mM formic acid. Error bars represent the standard deviation of three replicated analysis.

0

5

10

15

20

25


Rec

over

y (%

)

DEHPi

DEHPi + 3 mM PBA

DEHPi+ 25 mM TFPBA


191

3.3.3. Sample pH and acceptor phase composition The effect of sample pH on EME was investigated in the range of pH 3-8

using 20 mM phosphate buffer solutions. As shown in Figure S2, extraction recoveries were lower at pH 3 and 8 whereas comparable values were obtained for pH 4, 5, 6, and 7. The drop in extraction efficiency at pH 3 could be related to a reduced affinity of TFPBA to complex target analytes under strongly acidic conditions. The drop in extraction efficiency at pH 8 could be due to a partial negative ionization of the target analytes and, most likely, to the formation of anionic complexes with TFPBA [27]. Boronic acids can form neutral esters in non-polar solvents whereas they tend to form anionic boronate esters in water at basic pH [27]. At pH 8, the formation of anionic complexes could be favored over the formation of neutral complexes. The transport of these negatively charged molecules through the SLM was hindered by the direction of the applied voltage, and thus, the extraction efficiency decreased. Finally, pH 4 was selected as optimum value in terms of recoveries, and also considering the higher stability of the target analytes under acidic conditions [23].

Figure S2. Effect of sample pH on EME of catecholamines. Extraction conditions: concentration of analytes, 1 mg L-1; SLM, DEHPi with 25 mM TFPBA; applied voltage, 50 V; extraction time, 10 min; acceptor solution, 20 mM formic acid. Error bars represent the standard deviation of three replicated analysis.

0

5

10

15

20

25

30


Rec

over

y (%

)

pH 3 pH 4 pH 5

pH 6 pH 7 pH 8

Capítulo 3

192

The effect of acceptor solution composition on complexation-mediated EME was evaluated employing acidic conditions to maintain the positive ionization of catecholamines. To this end, solutions of 20 mM formic acid (pH=2.7), 200 mM formic acid (pH=2.2), 20 mM phosphate buffer (pH=2), and 10 mM hydrochloric acid (pH=2) were prepared and used as acceptor solution in different experiments. EME was performed from pH 4 sample solution, using DEHPi with 25 mM TFPBA in the SLM and an applied voltage of 50 V. After 10 min of extraction, comparable extraction efficiencies (data not shown) were obtained with the different experiments, showing a negligible effect of the acceptor solution composition on the EME. Finally, 20 mM formic acid was selected in subsequent experiments considering its compatibility with the UHPLC-MS/MS system used to evaluate the method.

3.3.4. Applied voltage and extraction time

The influence of applied voltage was studied from 0 to 100 V. In experiments at 0 V, catecholamines were not found in acceptor solution, and this supported that there were no passive diffusion of the model analytes in the current complexation-mediated EME systems. Thus, the use of voltage across the SLM was required to extract highly polar target analytes. The effect of voltage on complexation-mediated EME is shown in Figure 4a. As expected, the extraction recoveries increased with increasing voltage up to 100 V. However, system current also increased with the applied voltage, and the current exceeded 50 µA at 75 and 100 V. For urine samples, system current was also expected to exceed 50 µA at 75 and 100 V and it was checked to be under this value at 50 V. Finally, 50 V was chosen as optimum extraction voltage, compromising extraction recovery and EME system stability (current below 50 µA).


193

(a)

(b)

Figure 4. Effect of (a) applied voltage, and (b) extraction time. Extraction conditions: concentration of analytes, 1 mg L-1; sample pH, 4; SLM, DEHPi with 25 mM TFPBA; applied voltage, 50 V (if not indicated); extraction time, 10 min (if not indicated); acceptor solution, 20 mM formic acid. Error bars represent the standard deviation of three replicated analysis.

0

5

10

15

20

25

30

35

40


Rec

over

y (%

) 25 V

50 V

75 V

100 V

0

5

10

15

20

25

30

35

40

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90

Rec

over

y (%

)

Time (min)

Dopamine

Epinephrine

Norepinephrine

Capítulo 3

194

Finally, extraction time was investigated and the results are shown in Figure 4b. Reoveries increased as a function of time during the first 15 min of extraction, as expected. Longer extraction times did not improve extraction recoveries and, according to previous publications, this effect could be attributed to pH changes in the acceptor solution due to electrolysis [22,28]. The extraction time effect was also evaluated in a real urine sample diluted with 20 mM phosphate buffer pH 4 (volume ratio 1:1). As with the aqueous samples, no improvement in recoveries was observed after 15 min of extraction (data not shown). Therefore, 15 min was finally selected as optimum time for complexation-mediated EME of the catecholamines.

3.4. Extraction performance in urine samples under optimized conditions The final EME system was based on the following optimized conditions:

SLM, DEHPi with 25 mM TFPBA; sample pH, 4; acceptor solution, 20 mM formic acid; applied voltage, 50 V; and extraction time, 15 min. Under these conditions, recoveries were 10% for E, 15% for NE and 29% for DA when EME was performed from aqueous standards. However, when analyzing urine samples extraction recoveries decreased significantly as discussed in Section 3.5.

EME is known to provide excellent sample clean-up since the SLM forms a hydrophobic barrier between the sample and acceptor solution. Figure 5a shows HPLC-UV chromatograms before and after complexation-mediated EME of a diluted urine sample (volume ratio 1:1, urine:20 mM phosphate buffer pH 4) at a 1 mg L-1 spiking level. Although some peaks from the sample matrix are present after EME, differences between the two chromatograms are obvious, indicating a high level of sample clean-up. This indicated that even though the complexation reagent improved the mass transfer of the highly polar model analytes across the SLM, the selective nature of this complexation prevented the bulk matrix of the urine sample from entering the SLM. Additionally, the system current profile is illustrated in Figure 5b, showing that the complexation-mediated EME system was highly stable in contact with the diluted urine sample under optimized conditions.


195

(a)

(b)

Figure 5. EME performance in urine sample under optimized conditions: (a) HPLC-UV chromatograms showing sample clean-up, and (b) system current profile.

0

10

20

30

40

50

60

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Cur

rent

(µA

)

Time (min)

0

10

20

30

40

50

60

70

80

90

100

0 2 4 6 8 10 12 14

Sign

al (m

AU

)

Time (min)

Urine sample direct injection

EME extract

Dopamine

Norepinephrine

Epinephrine

Capítulo 3

196

3.5. Evaluation

Finally, the complexation-mediated EME concept was combined with UHPLC-MS/MS and evaluated for quantification of E and DA. The main purpose of this was to test if the new concept of complexation-mediated EME can provide reliable data. A complete validation was not considered at this stage. Quality analytical parameters were evaluated in pooled urine from three healthy volunteers. Standard addition calibration was used due to the matrix effects. To this end, pooled urine sample was diluted with 20 mM phosphate buffer of pH 4 (volume ratio 1:1) and calibration curves were constructed using standards of five concentration levels from 25 to 125 µg L-1. The content of E in the pooled urine sample was under the limit of detection (LOD) of the method whereas the content of DA was under the limit of quantification (LOQ). Correlation coefficient values (r) were 0.990 for E and 0.996 for DA. The Student’s t-test was applied to assess the linearity showing values of 11.91 (r=0.990, N=5) for E and to 8.20 (r=0.996, N=5) for DA, thus rejecting the null hypothesis of non-linear correlation for a 5% significance level and 3 degrees of freedom (t0.05,3=3.18) [29]. The repeatability of the method, expressed as coefficient of variation (CV), was determined by five consecutive extractions from diluted pooled urine sample spiked at a concentration level of 50 µg L-1. CV was 13% for both E and DA.

Extraction recoveries of the proposed procedure were found by the following strategy. First, diluted pooled urine sample was spiked at 50 µg L-1

with E and DA and subjected to EME. Then, EME was conducted from non-spiked diluted pooled urine and the final extract was spiked at 50 µg L-1. Signals obtained in both experiments were compared and, considering acceptor and sample solution volumes (i.e., 25 µL and 1 mL, respectively), extraction recoveries were calculated to (5.5±0.9)% for E and (15±2)% for DA (n=5). At this recovery level, LODs (S/N=3) were 5.0 and 1.8 µg L-1, and LOQs (S/N=10) were 16.5 and 6.0 µg L-1 for E and DA, respectively. Enrichment factors were 2.2 for E and 6.0 for DA. Although low enrichment factor were obtained, they could be further improved increasing sample and acceptor phases volume ratios.

Finally, diluted pooled urine sample was spiked at a known concentration level (i.e., 50 µg L-1) and analyzed by standard addition calibration using


197

standards of five concentration levels from 25 to 125 µg L-1. Relative recoveries were calculated as ratio between found and spiked concentrations being (91±26)% for E and (117±20)% for DA, where standard deviation values were calculated using the sXE (i.e., standard deviation of x-value estimated using regression line [29]).

4. Conclusions

In this work, complexation-mediated EME of highly polar basic drug substances was demonstrated for the first time using selected catecholamines as model analytes. Complexation in the bulk sample with water soluble PBA derivatives added to the sample, and complexation at the sample/SLM interface with water insoluble PBA derivatives added to the SLM were tested, and the latter concept appeared to be most efficient. Thus, complexation of the catecholamines with TFPBA at the sample/SLM interface was found to enhance the mass transfer across the SLM. Because the complexation reaction involved substances with two phenolic groups in ortho position only, the reaction was selective and therefore complexation-mediated EME appeared to be selective even from biological fluids. Thus, although the SLM permitted mass transfer of target analytes with -1>log P>-2, most bulk matrix components in human urine was unable to pass the SLM, and acceptable sample clean-up was achieved. Additionally, the current in the complexation-mediated EME system was easily controlled and kept below 50 µA, and therefore the system provided acceptable stability. The complexation-mediated EME concept was combined with UHPLC-MS/MS to develop a model application. Although the work presented in this paper is preliminary in nature, complexation-mediated EME showed potential and extraction of basic drugs with log P in the range -1 to -2 was demonstrated for the first time. Complexation-mediated EME should be investigated in more detail in the future. With this concept, analyte detection may be performed with instruments much more simple than mass spectrometry (as used in this initial work), and this may open new and very interesting future possibilities.

Capítulo 3

198

Acknowledgements

The authors would like to thank the Spanish Ministry of Science and Innovation (project n. CTQ2011-23968) and Generalitat Valenciana (Spain) (projects n. GVA/2014/096 and PROMETEO/2013/038) for the financial support. E. Fernández thanks Spanish Ministry of Education for her FPU grant (FPU13/03125) and mobility grant (EST15/00074).

Compliance with Ethical Standards

Informed consent was obtained from all individual participants included in the study. Urine samples were collected from healthy volunteers and randomized. Collection was performed in accordance with ethical standards and approved by the Director of School of Pharmacy (University of Oslo, Norway).

Conflict of Interest

The authors declare that they have no conflicts of interest.

References

1. Ghambarian M, Yamini Y, Esrafili A. Developments in hollow fiber based liquid-phase microextraction: principles and applications. Microchim Acta. 2012; 177:271-94.

2. Huang C, Jensen H, Seip KF, Gjelstad A, Pedersen-Bjergaard S. Mass transfer in electromembrane extraction-the link between theory and experiments. J Sep Sci. 2016; 39:188-97.

3. Fernández E, Vidal L. Liquid-phase microextraction techniques. In: Pena-Pereira F, editor. Miniaturization in sample preparation. Warsaw: De Gruyter Open; 2014. pp. 191-252.

4. Gjelstad A, Pedersen-Bjergaard S. Recent developments in electromembrane extraction. Anal Methods. 2013; 5:4549-57.


199

5. Marothu VK, Gorrepati M, Vusa R. Electromembrane extraction-a novel extraction technique for pharmaceutical, chemical, clinical and environmental analysis. J Chromatogr Sci. 2013; 51:619-31.

6. Gjelstad A, Rasmussen KE, Pedersen-Bjergaard S. Electrokinetic migration across artificial liquid membranes tuning the membrane chemistry to different types of drug substances. J Chromatogr A. 2006; 1124:29-34.

7. Nojavan S, Fakhari AR. Electro membrane extraction combined with capillary electrophoresis for the determination of amlodipine enantiomers in biological samples. J Sep Sci. 2010; 33:3231-38.

8. Fakhari AR, Tabani H, Nojavan S, Abedi H. Electromembrane extraction combined with cyclodextrin-modified capillary electrophoresis for the quantification of trimipramine enantiomers. Electrophoresis. 2012; 33:506-15.

9. Davarani SSH, Najarian AM, Nojavan S, Tabatabaei M-A. Electromembrane extraction combined with gas chromatography for quantification of tricyclic antidepressants in human body fluids. Anal Chim Acta. 2012; 725:51-6.

10. Hasheminasab KS, Fakhari AR. Development and application of carbon nanotubes assisted electromembrane extraction (CNTs/EME) for the determination of buprenorphine as a model of basic drugs from urine samples. Anal Chim Acta. 2013; 767:7580.

11. Ahmar H, Fakhari AR, Tabani H, Shahsavani A. Optimization of electromembrane extraction combined with differential pulse voltammetry using modified screen-printed electrode for the determination of sufentanil. Electrochim Acta. 2013; 96:117-23.

12. Huang C, Eibak LEE, Gjelstad A, Shen X, Trones R, Jensen H, Pedersen-Bjergaard S. Development of a flat membrane based device for electromembrane extraction: A new approach for exhaustive extraction of basic drugs from human plasma. J Chromatogr A. 2014; 1326:7-12.

13. Asl YA, Yamini Y, Seidi S, Amanzadeh H. Dynamic electromembrane extraction: Automated movement of donor and acceptor phases to improve extraction efficiency. J Chromatogr A. 2015; 1419:10-8.

Capítulo 3

200

14. Rouhollahi A, Kouchaki M, Seidi S. Electrically stimulated liquid phase microextraction combined with differential pulse voltammetry: a new and efficient design for in situ determination of clozapine from complicated matrices. RSC Adv. 2016; 6:12943-52.

15. Gjelstad A, Rasmussen KE, Pedersen-Bjergaard S. Electromembrane extraction of basic drugs from untreated human plasma and whole blood under physiological pH conditions. Anal Bioanal Chem. 2009; 393:921-8.

16. Eibak LEE, Gjelstad A, Rasmussen KE, Pedersen-Bjergaard S. Kinetic electro membrane extraction under stagnant conditions - fast isolation of drugs from untreated human plasma. J Chromatogr A. 2010; 1217:5050-6.

17. Jamt REG, Gjelstad A, Eibak LEE, Øiestad EL, Christophersen AS, Rasmussen KE, Pedersen-Bjergaard S. Electromembrane extraction of stimulating drugs from undiluted whole blood. J Chromatogr A. 2012; 1232:27-36.

18. Šlampová A, Kubáň P, Boček P. Electromembrane extraction using stabilized constant d.c. electric current - a simple tool for improvement of extraction performance. J Chromatogr A. 2012; 1234:32-7.

19. Kjelsen IJØ, Gjelstad A, Rasmussen KE, Pedersen-Bjergaard S. Low-voltage electromembrane extraction of basic drugs from biological samples. J Chromatogr A. 2008; 1180:1-9.

20. Seip KF, Faizi M, Vergel C, Gjelstad A, Pedersen-Bjergaard S. Stability and efficiency of supported liquid membranes in electromembrane extraction-a link to solvent properties. Anal Bioanal Chem. 2014; 406:2151-61.

21. Huang C, Seip KF, Gjelstad A, Pedersen-Bjergaard S. Electromembrane extraction for pharmaceutical and biomedical analysis - Quo vadis. J Pharm Biomed Anal. 2015; 113:97-107.

22. Huang C, Seip KF, Gjelstad A, Pedersen-Bjergaard S. Electromembrane extraction of polar basic drugs from plasma with pure bis(2-ethylhexyl) phosphite as supported liquid membrane. Anal Chim Acta. 2016; 934:80-7.

23. Bicker J, Fortuna A, Alves G, Falcão A. Liquid chromatographic methods for the quantification of catecholamines and their metabolites in several biological samples - A review. Anal Chim Acta. 2013; 768:12-34.


201

24. Paugam M-F, Valencia LS, Boggess B, Smith BD. Selective dopamine transport using a crown boronic acid. J Am Chem Soc. 1994; 116:11203-4.

25. Takeuchi M, Koumoto K, Goto M, Shinkai S. Efficient glucoside extraction mediated by a boronic acid with an intramolecular quaternary ammonium ion. Tetrahedron. 1996; 52:12931-40.

26. Di Luccio M, Smith BD, Kida T, Borges CP, Alves TLM. Separation of fructose from a mixture of sugars using supported liquid membranes. J Memb Sci. 2000; 174:217-24.

27. Kanamori T, Funatsu T, Tsunoda M. Determination of catecholamines and related compounds in mouse urine using column-switching HPLC. Analyst. 2016; 141:2568-73.

28. Kubán P, Boček, P. The effects of electrolysis on operational solutions in electromembrane extraction: The role of acceptor solution. J Chromatogr A. 2015; 1398:11-9.

29. Miller JN, Miller JC, Statistics and chemometrics for analytical chemistry. fifth ed. London: Pearson Prentice Hall; 2005.

Capítulo 4


205

Screen-printed electrode-based electrochemical detector coupled with in-situ ionic-liquid-assisted dispersive liquid-liquid

microextraction for determination of 2,4,6-trinitrotoluene

Elena Fernándeza, Lorena Vidala, Jesús Iniestab, Jonathan P. Mettersc, Craig E. Banksc and Antonio Canalsa

a Departamento de Química Analítica, Nutrición y Bromatología e Instituto Universitario de Materiales, Universidad de Alicante, P.O. Box 99, 03080 Alicante, Spain. b Departamento de Química Física e Instituto Universitario de Electroquímica, Universidad de Alicante P.O. Box 99, 03080 Alicante, Spain. c Faculty of Science and Engineering, Chemistry and Environmental Science, Manchester Metropolitan University, Chester Street, Manchester M1 5GD, UK

Received: 31 July 2013 / Revised: 25 September 2013 / Accepted: 2 October 2013 / Published online: 19 November 2013

Abstract

A novel method is reported, whereby screen-printed electrodes (SPELs) are combined with dispersive liquid-liquid microextraction. In-situ ionic liquid (IL) formation was used as an extractant phase in the microextraction technique and proved to be a simple, fast and inexpensive analytical method. This approach uses miniaturized systems both in sample preparation and in the detection stage, helping to develop environmentally friendly analytical methods and portable devices to enable rapid and onsite measurement. The microextraction method is based on a simple metathesis reaction, in which a water-immiscible IL (1-hexyl-3-methylimidazolium bis[(trifluoromethyl)sulfonyl]imide, [Hmim][NTf2]) is formed from a water-

Anal Bioanal Chem (2014) 406:2197-2204 DOI 10.1007/s00216-013-7415-y

Capítulo 4

206

miscible IL (1-hexyl-3-methylimidazolium chloride, [Hmim][Cl]) and an ion-exchange reagent (lithium bis[(trifluoromethyl)sulfonyl]imide, LiNTf2) in sample solutions. The explosive 2,4,6-trinitrotoluene (TNT) was used as a model analyte to develop the method. The electrochemical behavior of TNT in [Hmim][NTf2] has been studied in SPELs. The extraction method was first optimized by use of a two-step multivariate optimization strategy, using Plackett-Burman and central composite designs. The method was then evaluated under optimum conditions and a good level of linearity was obtained, with a correlation coefficient of 0.9990. Limits of detection and quantification were 7 μg L-1 and 9 μg L-1, respectively. The repeatability of the proposed method was evaluated at two different spiking levels (20 and 50 μg L-1), and coefficients of variation of 7% and 5% (n=5) were obtained. Tap water and industrial wastewater were selected as real-world water samples to assess the applicability of the method.

Keywords: Liquid phase microextraction, dispersive liquid-liquid microextraction, ionic liquid, screen printed electrodes, 2,4,6-trinitrotoluene, water samples.

1. Introduction

Miniaturization of both analytical methods and instrumentation has become very popular in recent years, and many efforts have focused on performing chemical analysis at a reduced scale. Miniaturization of sample preparation methods has substantially increased, with the development of many solid-phase and liquid-phase microextraction techniques [1,2]. Compared with traditional liquid-liquid extraction techniques, liquid-phase microextraction (LPME) offers simplicity, ease of handling, minimal sample and solvent consumption, and an important reduction in residues generated. Since its appearance in the nineties, several LPME techniques have been developed, with single-drop microextraction, hollow fiber liquid-phase microextraction and dispersive liquid-liquid microextraction [3] (DLLME) the most commonly used. Organic solvents have traditionally been used as extractants in LPME techniques, but use of ionic liquids (ILs) has recently attracted interest as a


207

promising alternative [4]. ILs are melted salts at room temperature and have unique properties; in particular, they are highly thermally and chemically stable, with negligible vapor pressure, tunable viscosity, electrolytic conductivity, a wide electrochemical window, and good extractability for organic compounds and metal ions [5]. Use of ILs has helped to overcome problems associated with LPME techniques using organic solvents [4] and has enabled the development of new methods, including temperature-controlled IL dispersive liquid-liquid microextraction [6], and in-situ IL formation dispersive liquid-liquid microextraction (in-situ IL-DLLME) [7,8]. During in-situ IL-DLLME the extractant phase is formed into the sample solution via a metathesis reaction between a water-miscible IL and an ion-exchange reagent, forming a water-immiscible IL. Homogeneously dispersed fine drops of the extractant phase are generated, and high enrichment factors are obtained with low extraction times because of the high contact surface between phases. Dispersion of the IL takes place via metathesis reaction, and a disperser agent is not needed; this avoids competition with the IL, which would reduce extraction efficiency. Moreover, this process does not need the additional devices, for example vortex or ultrasound bath, which have been used to assist IL-DLLME [9, 10].

Most LPME procedures are followed by chromatographic separation, in either liquid or gas modalities, coupled with different detection systems (UV-visible, inductively coupled plasma optical emission spectrometry, or mass spectrometry, among others). Most of these detection systems are slow, expensive and bulky, so analytical instrumentation used for detection has not achieved miniaturization to the same extent as sample preparation methods, miniaturized forms of which are more widely used. Furthermore, ILs have some disadvantages when chromatographic techniques are used. For example, special devices are needed when ILs are injected in gas chromatography, because of their high boiling points [11,12], and shorter column life and resolution problems are challenges for liquid chromatography with ILs. Accordingly, electrochemical sensors are regarded as an attractive option for use in detection methods. Recent advances in microfabrication and screen-printing technology have enabled the development of miniaturized and easy-to-use electrochemical systems for rapid and decentralized onsite measurements. Screen-printed electrodes [13] (SPELs)

Capítulo 4

208

are inexpensive, mass-produced, disposable devices, which are ideal for low volume sample analysis. The versatility and ease of modification of these electrodes are advantageous for improving their performance and adapting them to specific analytes.

As far as we are aware, this is the first report of an analytical method in which SPELs are used as electrochemical detectors for LPME. ILs have intrinsic conductivity, a wide electrochemical window, and thermal stability, and have therefore been recognized as ideal alternative electrolytes in electrochemical devices [14]. Taking into account the electrochemical properties of ILs and the low volume of IL-phase formed during in-situ IL-DLLME (10-20 μL), SPELs seem to be ideal and perfectly compatible candidates for analyzing IL drop after microextraction without any further modification. The explosive 2,4,6-trinitrotoluene (TNT) was used as a model analyte to develop the proposed method, because its electrochemical behavior has been widely characterized. Additionally, nitroaromatic explosives, including TNT, have been the subject of increasing interest in recent years. Concerns about terrorist activity have led to intensification of securitymeasures in airports and public buildings, creating a demand for highly sensitive analytical methods to detect these compounds at trace levels. Furthermore, their presence in surrounding soils, waterways and reservoirs must be monitored in areas where they are produced, stored or detonated. The mutagenic and toxic properties of nitroaromatic explosives make their presence in the environment dangerous [15]. Many analytical methods using gas [16], liquid [17] and micellar electrokinetic [18] chromatography and immunoassay techniques [19] have been developed to determine TNT and other related explosives in environmental samples. However, the inherent redox activity of TNT makes electrochemical sensors a very suitable alternative. Electrochemistry offers simplicity, rapid response, and low-cost instrumentation with portable options, and meets the analytical requirements of sensitivity and reproducibility.

Screen-printed electrodes have previously been used in electrochemical devices for TNT determination. Hydrogel-coated SPELs have been used to detect thermally-desorbed TNT from an integrated preconcentration system for both solid and liquid samples [20]. An electrochemically pre-anodized Nafion-coated screen-printed carbon electrode (SPCE) has been used in a disposable sensor developed to determine different nitroaromatic compounds [21]. With this system, TNT can be


209

detected in spiked lake water at a concentration of 30 μmol L-1 (6.8 μg mL-1). A wearable textile-based screen-printed electrochemical sensor, which is highly suitable for monitoring the surroundings of the wearer, has been tested for detection of nitroaromatic explosives both in gas and liquid phase, obtaining a limit of detection (LOD) of approximately 1 μg mL-1 for TNT in water [22]. Recently, a very simple and low-cost sensor based on unmodified SPCEs has been described to determine TNT and 2,4-dinitrotoluene in aqueous solutions, obtaining LODs as low as 0.4 μmol L-1 (90 μg L-1) and 0.7 μmol L-1 (73 μg L-1), respectively [23].

The research presented here combines the advantages of LPME techniques with the benefits that SPELs offer as electrochemical sensors. The resulting novel method uses miniaturized systems in both sample preparation and detection. In addition, the use of an IL as the extractant phase not only provides the advantage of good extractability of organic compounds, but also provides the electrolyte behavior needed for detection. The proposed method has been optimized by use of a multivariate optimization strategy, and its ability to determine TNT in real-world water samples has been established.

2. Experimental

2.1. Reagents and real world water samples

A TNT standard of 1000 mg L-1 in acetonitrile was obtained from LGC Standards (Warsaw, Poland). Stock solution of TNT (10 mg L-1) in HPLC grade acetonitrile from Sigma-Aldrich (Seelze, Germany) was prepared and stored in the dark at 4 °C. Working solutions were prepared daily by appropriate dilution of this stock solution in ultrapure water from a water purification system (MiliQ Biocel A10) supplied by Millipore (Bilerica, MA, USA). ILs 1-hexyl-3-methylimidazolium bis[(trifluoromethyl)sulfonyl]imide ([Hmim][NTf2]) (98%) and 1-hexyl-3-methylimidazolium chloride ([Hmim][Cl]) (98%) were purchased from Iolitec (Heilbronn, Germany). Lithium bis[(trifluoromethyl)sulfonyl]imide (LiNTf2) salt was supplied by Sigma-Aldrich (St. Louis, MO, USA), and reactive grade NaCl by ACS Scharlau (Barcelona, Spain).

Capítulo 4

210

Tap water from a drinking water treatment plant in Seville (Spain) and industrial wastewater from Ourense (Spain) were used as real-world water samples. The industrial wastewater contained a chemical oxygen demand of 1004 mg O2 L-1, a biochemical oxygen demand of 278 mg O2 L-1, and 429 mg L-1 suspended solids. Samples were collected in amber glass containers and stored in the dark at 4 °C. Tap water was used without any further pretreatment. The wastewater was filtered with a 0.22 μm pore-size nylon filter before use. The real-world water samples were previously analyzed, and the target analyte was not detected.

2.2. In-situ IL-DLLME procedure

Under optimum conditions, 32 mg [Hmim][Cl] was placed in a conical-bottom glass tube and dissolved in 11 mL aqueous standard or 11 mL sample solution. An equimolar quantity of LiNTf2 (45.3 mg) relative to [Hmim][Cl] was added, and a cloudy solution immediately formed. The mixture was manually shaken for 0.5 min. To accelerate phase separation, the tube was then placed in an ice bath for 3 min. Next, the phases were separated by centrifugation for 5 min at 4000 rpm. The aqueous phase was removed with a glass pipette, and 15 μL of the formed IL phase (i.e., [Hmim][NTf2]) was withdrawn with a syringe. Finally, this 15 μL IL phase was deposited on the screen-printed graphite electrode (SPGE) surface for electrochemical detection. The procedure is described in Figure 1.


211

Figure 1. In-situ IL-DLLME coupled with SPGE.

2.3. Electrochemical analysis

A μ-Autolab III potentiostat/galvanostat from Eco Chemie (Utrecht, The Netherlands) controlled by Autolab GPEs software version 4.9 for Windows XP was used for electrochemical experiments. All measurements were performed by use of three-electrode configuration SPGEs from Kanichi Research Services (Manchester, UK). SPGEs were manufactured as described elsewhere [24]. The working electrode, of 3.1 mm diameter, and the counter

Centrifugation

Sample addition

[Hmim][Cl] LiNTf2 addition

Ice bath

Shaking

[Hmim][NTf2] enriched with the analyte

Removing aqueous phase

Deposition of the IL drop on the electrode surface

Capítulo 4

212

electrode were made from a graphite ink. A pseudo Ag/AgCl was the reference electrode. Connectors for the electrochemical connection of the SPGEs were also obtained from Kanichi Research Services Ltd. SPGEs were used without any pretreatment or modification of the working electrode surface, and a new SPGE was used for each experiment.

Differential pulse voltammetry (DPV) was selected as the electroanalytical technique. A variety of DPV variables were optimized by use of a one-at-a-time strategy, establishing the following optimum conditions: 100 mV modulation amplitude; 10 mV step potential; 0.05 s modulation time; and 0.5 s interval time. Pure N2 from Air Liquide (Madrid, Spain) flowed for 20 minutes before DPV experiments and was maintained during measurements. The signal corresponding to oxygen embedded in the IL [25] appears at the same reduction potential as the analyte, meaning purging with N2 was necessary to ensure a deoxygenated atmosphere in which the analyte could be detected at low concentrations. All electrochemical measurements were performed at room temperature.


The current peak of the first cathodic wave of TNT at -0.80 V vs. pseudo Ag/AgCl was used to identify and quantify the analyte in order to evaluate the developed method. A multivariate optimization strategy was performed to determine the optimum conditions for the microextraction method. Statgraphics statistical computer package “Statgraphics Plus 5.1.” (Warrenton, VA, USA) was used to construct the experimental design matrices and evaluate the results.


213


3.1. Multivariate optimization

3.1.1. Screening step

Plackett-Burman design is a two-level fractional factorial design for studying k=N-1 variables in N runs, where N is a multiple of four [26]. Fractional factorial designs are very useful in the first steps of a project, when many variables are initially investigated but only a few have important effects. The Plackett-Burman design ignores interaction between variables so the main effects can be calculated by means of a reduced number of experiments, enabling more economical experimentation. A saturated Plackett-Burman design was used to construct the matrix of experiments, including 11 variables: eight real variables and three dummy variables. The effects of the dummy variables were used to evaluate experimental error [27, 28].

On the basis of the literature [7] and of the previous experience of the research group [29], the eight real experimental variables selected at two levels were: amount of [Hmim][Cl], sample volume, molar ratio between [Hmim][Cl] and the salt LiNTf2, ionic strength, extraction time, centrifugation speed, centrifugation time, and purge time with N2 before electrochemical measurements. Table 1 shows the experimental variables and levels considered in the Plackett-Burman design. A total of twelve experiments were performed, using aqueous standards of 100 μg L-1.

Capítulo 4

214

Table 1. Experimental variables and levels of the Plackett-Burman design.

The data obtained were evaluated by use of an ANOVA test, and the

results were displayed in the Pareto chart shown in Figure S1(a). The length of each bar is proportional to the effect of the corresponding variable, and the effects that exceed the reference vertical line can be regarded as significant with a 95% probability. According to Figure S1(a), only the amount of [Hmim][Cl] was a statistically significant variable with 95% probability, having a negative effect. This negative effect is in agreement with the fact that if a smaller amount of [Hmim][Cl] is used, a smaller volume of IL-phase is formed in the microextraction procedure, and a higher concentration of analyte is therefore obtained in the extraction phase.

Sample volume is the second most important variable; its positive effect is non-significant, but is much larger than the effect of purge time. Our previous experience indicated that sample volume is an important variable in microextraction techniques. In general, greater sample volume involves a greater amount of analyte and therefore increases the response. For this reason, the amount of [Hmim][Cl] and the sample volume were selected as the main variables affecting the response of the system, and were investigated during the optimization of significant variables. The other six real variables with non-significant effects were fixed at the most experimentally convenient level, specifically: stoichiometric molar ratio between [Hmim][Cl] and LiNTf2;

Variable Level

Low (-1) High (+1) Amount of [Hmim][Cl] (mg) 34 68

Sample volume (mL) 5 11

Ionic strength (NaCl concentration, %, w/v) 0 10

Molar ratio [Hmim][Cl]:LiNTf2 1:1 1:3

Extraction time (min) 0.5 1

Centrifugation time (min) 5 10

Centrifugation speed (rpm) 2000 4000

Purge time (min) 20 30


215

ionic strength: 0% NaCl; extraction time: 0.5 min; centrifugation speed: 4000 rpm; centrifugation time: 5 min; and purge time: 20 min.

(a)

(b)

Figure S1. Pareto charts of the multivariate optimization. (a) Plackett-Burman design. (b) Central composite design. AA and BB are the quadratic effects of amount of [Hmim][Cl] and sample volume, respectively. AB is the interaction effect between the two variables.

Amount of [Hmim][Cl] Sample volume

Purge time Dummy 1

Molar ratio [Hmim][Cl]:LiNTf2 Dummy 2

Centrifugation time Ionic strength

Centrifugation speed Dummy 3

Extraction time

Standarized effect 0 0.4 0.8 1.2 1.6 2 2.4

Standarized effect

Diagrama de Pareto para Var_1

0 0.4 0.8 1.2 1.6 2 2.4

D:Factor_D

K:Factor_K

F:Factor_F

E:Factor_E

B:Factor_B

J:Factor_J

G:Factor_G

I:Factor_I

H:Factor_H

C:Factor_C

A:Factor_A+

-Sample volume

Purge time

Molar ratio [Hmim][Cl]:LiNTf2

Centrifugation time

Centrifugation speed

Extraction time

Ionic strength

Dummy 2

Amount of [Hmim][Cl]

Dummy 3

Dummy 1

Standarized ef fect

Standarized effect

Diagrama de Pareto para Var_1

0 0.4 0.8 1.2 1.6 2 2.4

D:Factor_D

K:Factor_K

F:Factor_F

E:Factor_E

B:Factor_B

J:Factor_J

G:Factor_G

I:Factor_I

H:Factor_H

C:Factor_C

A:Factor_A+

-Sample volume

Purge time

Molar ratio [Hmim][Cl]:LiNTf2

Centrifugation time

Centrifugation speed

Extraction time

Ionic strength

Dummy 2


Dummy 3

Dummy 1

Standarized ef fect

Amount of [Hmim][Cl] (A)

Sample volume (B)

AA

BB

AB

0 1 2 Standarized effect

3 4 5

Amount of [Hmim][Cl] (A)

AA

Sample volume (B)

BB

AB

Standarized ef fectStandarized ef f ect

Diagrama de Pareto Estandarizada para Var_1

0 1 2 3 4 5

AB

BB

B:Factor_B

AA

A:Factor_A +

-

Capítulo 4

216

3.1.2. Optimization of significant variables

Central composite design (CCD) was used in this optimization step. CCD combines a two-level full factorial design (2k) with 2k star points, where k is the number of variables being optimized, and one point at the center of the experimental region, that can be run n times. To ensure the rotatability of the

model, star points were set at α=√2k4=1.41, and the central point was repeated

five times to provide an orthogonal design [26]. CCD was used to evaluate and optimize main effects, interaction effects and quadratic effects of the two significant variables. Table 2 shows the low and high levels, and the central and star points of the two variables in the optimization step. Thirteen experiments were performed, using aqueous standards of 100 μg L-1.

Table 2. Variables, low and high levels, central and star points used in CCD.

The data obtained were also evaluated by use of an ANOVA test and the results were displayed in the Pareto chart shown in Figure S1(b). As can be seen, both the amount of [Hmim][Cl] and the sample volume were significant variables, with a 95% probability, confirming the predicted importance of sample volume effect. One of the quadratic effects was also significant, assuming the curvature of the system and fitting the second-grade polynomial model proposed. The response surface obtained by use of the CCD is shown in Figure 2.

Variable Level

Star points (1.41)

Low (-1) Central (0) High (+1) - +

Amount of [Hmim][Cl] (mg) 40 60 80 32 88

Sample volume (mL) 4 7 10 3 11


217

Figure 2. Response surface of CCD.

The surface graph shows a pronounced increase in the analytical signal as the amount of [Hmim][Cl] decreases and the sample volume increases. In summary, the results obtained from the optimization process led to the following experimental conditions: amount of [Hmim][Cl], 32 mg; sample volume, 11 mL; molar ratio [Hmim][Cl]:LiNTf2, 1:1; ionic strength, 0% NaCl; extraction time, 0.5 min; centrifugation speed, 4000 rpm; centrifugation time, 5 min; and purge time before electrochemical measurements, 20 min.

3.2. Electrochemical study of TNT in [Hmim][NTf2] at SPGEs

DPV of a blank and of four TNT standards of 10, 30, 50 and 70 mg L-1, prepared in commercial [Hmim][NTf2], was performed using SPGEs to study the electrochemical behavior of the analyte. According to previous studies [30,31], using ILs as electrolytes usually leads to three consecutive reduction peaks of TNT, believed to correspond to each nitro group of the aromatic ring. Figure 3 shows DPV curves obtained in this study.

Cur

rent

pea

k (µ

A)

Var_1

1.6

2.4

3.2

4.0

4.8

5.6

6.4

7.2

Superficie de Respuesta Estimada

-1.40

1.4-1.4

0

1.4

0

2

4

6

8

10

Cu

rre

ntp

ea

k(µ

A)


Sample volumeSample volume

Amount of [Hmim][Cl] -1.4

-1.4 0 1.4

1.4

0

10

8

6

4

2

0

Capítulo 4

218

Figure 3. DPV curves of a blank and four TNT standards prepared in commercial [Hmim][NTf2].

A well-defined cathodic peak appears at -0.80 V, corresponding to the reduction of one of the three nitro groups in the molecule; the two peaks corresponding to the remaining nitro groups cannot be clearly distinguished. The current peak at -0.80 V had a good linearity, between 10 and 70 mg L-1, with a correlation coefficient (r) of 0.996. The repeatability of the electrochemical response was evaluated for five repeated analyses of the 70 mg L-1 standard, and a coefficient of variation (CV) of 2% was found. Next, the electrochemical behavior of TNT in [Hmim][NTf2] generated in-situ was studied. As shown in Figure 4, the reduction peaks of the three nitro groups are perfectly well defined and, as in commercial [Hmim][NTf2], an identical peak appears at -0.80 V. However, the reductive peaks at -1.06 V and -1.27 V are not well defined at lower concentrations, and therefore the current peak at -0.80 V was chosen for evaluation of the proposed method.

blank 10 g L-1

30 g L-1

50 g L-1

70 g L-1

-2.0 -1.5 -1.0 -0.5 0.0

-2.5

-2.0

-1.5

-1.0

-0.5

0.0

Cur

rent

(A

)

Potential (V) vs Ag/AgCl

-2.0 -1.6 -1.2 -0.8 -0.4 0.0

-2.0

-1.5

-1.0

-0.5

0.0C

urre

nt (

A)


blank 80 g L-1


219

Figure 4. DPV curves of a blank and a TNT standard of 80 μg L-1 in ultrapure water after in-situ IL-DLLME under optimum conditions.

3.3. Analytical figures of merit of the proposed method

Quality variables of the proposed method were evaluated. A concentration range from 10 to 100 μg L-1 was studied and the linear range was established to be from 10 to 80 μg L-1.The resulting calibration curve revealed a high level of linearity, with a correlation coefficient (r) of 0.9990 (N=4). The sensitivity of the instrumental measurements estimated from the slope of the calibration curve was (0.0112±0.0004) μA μg-1 L. The repeatability of the proposed method, expressed as CV, was evaluated for two spiking levels (20 and 50 μg L-1) by extracting five consecutive aqueous standards, and CV values were found to be 7% and 5%, respectively. The enrichment factor of the proposed procedure was 300, defined as the ratio of Co/Ca, where Co is the concentration of analytes in the IL phase after extraction and Ca is the original concentration of analytes in the aqueous phase. The LOD and the limit of quantification (LOQ) were estimated by using the mean signal of the blank (n=three replicates) at -0.80 V plus three or ten times its standard deviation. The LOD was found to be 7 μg L-1, and the LOQ was 9 μg L-1. The LOD of the

-2.0 -1.6 -1.2 -0.8 -0.4 0.0

-2.0

-1.5

-1.0

-0.5

0.0

Cur

rent

(A

)


blank 80 g L-1

Capítulo 4

220

developed method is therefore equal to or lower than other recently reported detection limits for the electrochemical determination of TNT obtained from more expensive and complex electrodes, using carbon nanomaterials, metallic nanoparticles or assembly procedures (Table 3). In addition, our approach combines a simple sample preparation step with unmodified, inexpensive SPGEs, thereby providing a lower LOD than those reported for other SPELs [20-23] (Table 3). Therefore, the developed method has unique benefits.

3.4. Real-world water sample analysis

The ability of the proposed method to determine TNT in real-world water samples was evaluated. Three replicated analyses of both tap water and wastewater were performed at a 40 μg L-1 spiking level. Wastewater was filtered with a 0.22 μm nylon filter after being spiked. It should be noted that in previous analyses none of the selected water samples had detectable initial TNT concentrations. Relative recoveries were calculated as the ratio of the signals obtained for real and ultrapure water samples spiked at the same concentration level. Relative recovery for tap water samples was found to be 114% with a CV value of 16%, whereas for wastewater samples the relative recovery was 109% with 18% CV. From these results, it can be concluded that matrix effects were not significant in TNT analysis of the two selected water samples. Therefore, the developed method can be successfully applied to both clean and complex water sample matrices.


221

Table 3. Characteristics of some electrochemical methods developed to determine TNT and other nitroaromatic explosives.

Electrode Analytical technique Analytes Real samples and spiking

levels in parentheses TNT LOD Ref.

SPGE DPV TNT Tap water and industrial wastewater (40 µg L-1) 7 µg L-1 This

work

Hydrogel-coated SPEL SWV TNT - - [20]

Nafion-coated SPCE SWV TNT and other nitroaromatic explosives

Lake water (30 µM; 6.8 µg mL-1) - [21]

Textile-based SPEL SWV TNT and 2,4-dinitrotoluene - 1 μg mL-1 [22]

SPCE CV TNT and 2,4-dinitrotoluene - 0.4 μM

(90 µg L-1) [23]

Mesoporous SiO2 modified GCE LSV TNT and other

nitroaromatic explosives - < 1.8 nM (4 µg L-1) [32]

GCE functionalized with OMC AdsSV TNT and other

nitroaromatic explosives - 0.2 μg L-1 [33]

SWV, square-wave voltammetry; CV, cyclic voltammetry; GCE, glassy carbon electrode; LSV, linear sweep voltammetry; OMC, ordered mesoporous carbon; AdsSV, adsorptive stripping voltammetry.

Capítulo 4

222

Table 3 (cont.). Characteristics of some electrochemical methods developed to determine TNT and other nitroaromatic explosives.

Electrode Analytical technique Analytes Real samples and spiking

levels in parentheses TNT LOD Ref.

GCE modified with MWCNTs AdsSV TNT Sea water

(100-1000 µg L-1) 0.6 μg L-1 [34]

GCE modified with Cu nanoparticles and SWCNTs

AdsSV TNT and other nitroaromatic explosives

River and tap water, and soil (50-200 µg L-1) 1 μg L-1 [35]

GCE modified by LbL self-assembled of

{MSU/PDDA}n films. DPV TNT and other

nitroaromatic explosives - 1.5 nM (3.4 μg L-1) [36]

Amalgam Hg/Au electrode Microchip FIA-ED TNT - 7 μg L-1 [37]

Electrochemically activated CFMEs SWV TNT and other

nitroaromatic explosives

Tap and ground water (2 mg L-1);

soil samples (200 mg g-1) 0.03 μg mL-1 [38]

BDD SWV TNT Sea water (20-400 µg L-1) 10 μg L-1 [39]

Graphene nanoribbon-modified GCE, graphene nanosheet-modified GCE

and bare GCE

DPV TNT Sea water (4-20 mg L-1)

0.140-0.520 mg L-1 [40]

GCE, glassy carbon electrode; MWCNTs, multi-wall carbon nanotubes; AdsSV, adsorptive stripping voltammetry; SWCNTs single-wallet carbon nanotubes; LbL, layer-by-layer; MSU, mesoporous SiO2; PDDA, poly(diallyldimethylammonium) chloride; FIA-ED, flow injection analysis with electrochemical detection; CFMEs, carbon fiber microelectrodes; SWV, square wave voltammetry; BDD, boron doped diamond.


223

4. Conclusions

Screen-printed electrode-based electrochemical detection has been successfully combined with LPME for the first time. The proposed analytical method uses miniaturized systems both in sample preparation and in the detection stage, and therefore has the advantage of avoiding expensive and bulky or immovable instrumentation. In-situ IL formation in the microextraction method avoids the use of harmful and toxic organic compounds as extractant and disperser solvents. The incorporation of a simple and fast sample preparation step before the electrochemical measurement by use of low-cost and disposable SPGEs has enabled a lower LOD than has been reported for other SPELs. The multivariate optimization strategy used here enabled us to rapidly and economically establish the optimum conditions for the main experimental variables involved in the sample preparation, thus providing complete information. Finally, the results prove the ability of the proposed method to determine TNT at trace levels in real-world water samples.

Although the use of a nitrogen purge, ice bath and the centrifuge limits portability, this method is a step forward in the development of portable and economical systems available to any laboratory.

Acknowledgments

The authors would like to thank the Spanish Ministry of Science and Innovation (project n. CTQ2011-23968), Generalitat Valenciana (Spain) (projects n. ACOMP/2013/072 and PROMETEO/2012/038) and Universidad de Alicante (Spain) (project n. GRE12-45) for the financial support. E.F. also thanks Generalitat Valenciana for her fellowship.

References

1. Risticevic S, Niri VH, Vuckovic D, Pawliszyn J (2009) Anal Bioanal Chem 393:781-795.

2. Pena-Pereira F, Lavilla I, Bendicho C (2010) Trends Anal Chem 29: 617-628.

Capítulo 4

224

3. Rezaee M, Assadi Y, Milani Hosseini M-R, Aghaee E, Ahmadi F, Berijani S (2006) J Chromatogr A 1116:1-9.

4. Aguilera-Herrador E, Lucena R, Cárdenas S, Valcárcel M (2010) Trends Anal Chem 29:602-616.

5. Sun P, Armstrong DW (2010) Anal Chim Acta 661:1-16. 6. Zhou Q, Bai H, Xie G, Xiao J (2008) J Chromatogr A 1177:43-49. 7. Baghdadi M, Shemirani F (2009) Anal Chim Acta 634:186-191. 8. Yao C, Anderson JL (2009) Anal Bioanal Chem 395:1491-1502. 9. Sun X-M, Sun Y, Wu L-W, Jiang C-Z, Yu X, Gao Y, Wang L-Y, Song D-

Q (2012) Anal Methods 4:2074-2080. 10. Zhang Y, Lee HK (2012) Anal Chim Acta 750:120-126. 11. Aguilera-Herrador E, Lucena R, Cárdenas S, Valcárcel M (2008) Anal

Chem 80:793-800. 12. Chisvert A, Román IP, Vidal L, Canals A (2009) J Chromatogr A

1216:1290-1295. 13. Metters JP, Kadara RO, Banks CE (2011) Analyst 136:1067-1076. 14. Buzzeo MC, Evans RG, Compton RG (2004) ChemPhysChem 5:1106-

1120. 15. Pichtel J (2012) Appl Environ Soil Sci 2012:1-33. 16. Cortada C, Vidal L, Canals A (2011) Talanta 85:2546-2552. 17. Sun Q, Chen Z, Yuan D, Yuan D, Yu C-P, Mallavarapu M, Naidu R

(2011) Chromatographia 73:631-637. 18. Giordano BC, Burgi DS, Collins GE (2010) J Chromatogr A 1217: 4487-

4493. 19. Goldman ER, Anderson GP, Lebedev N, Lingerfelt BM, Winter PT,

Patterson CH, Mauro JM (2003) Anal Bioanal Chem 375:471-475. 20. Cizek K, Prior C, Thammakhet C, GalikM, Linker K, Tsui R, Cagan A,

Wake J, La Belle J,Wang J (2010) Anal Chim Acta 661:117-121. 21. Chen J-C, Shih J-L, Liu C-H, Liu C-H, Kuo M-Y, Zen J-M (2006) Anal

Chem 78:3752-3757. 22. Chuang M-C, Windmiller JR, Santhosh P, Santhosh P, Ramirez GV, Galik

M, Chou T-Y, Wang J (2010) Electroanalysis 22:2511-2518. 23. Caygill JS, Collyer SD, Holmes JL, Davis F, Higson S (2013) Analyst

138:346-352.


225

24. Hallam PM, Kampouris DK, Kadara RO, Banks CE (2010) Analyst 135:1947-1952.

25. Lee J, Murugappan K, Arrigan DWM, Silvester DS (2013) Electrochim Acta 101:158-168.

26. Montgomery DC (2009) Design and Analysis of Experiments, 7th edn. Wiley, New Jersey.

27. Heyden YV, Hartmann C, Massart DL, Michel L, Kiechle P, Erni F (1995) Anal Chim Acta 316:15-26.

28. Fabre H, Mesplet N (2000) J Chromatogr A 897:329-338. 29. Cortada C, Vidal L, Canals A (2011) J Chromatogr A 1218:17-22. 30. Rabenecker P, Pinkwart K (2009) Propellants Explos Pyrotech 34:274-

279. 31. Xiao C, Rehman A, Zeng X (2012) Anal Chem 84:1416-1424. 32. Zhang H-X, Cao A-M, Hu J-S, Wan L-J, Lee S-T (2006) Anal Chem

78:1967-1971. 33. Zang J, Guo CX, Hu F, Yu L, Li CM (2011) Anal Chim Acta 683:187-

191. 34. Wang J, Hocevar SB, Ogorevc B (2004) Electrochem Commun 6:176-

179. 35. Hrapovic S, Majid E, Liu Y, Male K, Luong J (2006) Anal Chem 78:

5504-5512. 36. Shi G, Qu Y, Zhai Y, Liu Y, Sun Z, Yang J, Jin L (2007) Electrochem

Commun 9:1719-1724. 37. Wang J, Pumera M (2006) Talanta 69:984-987. 38. Agüí L, Vega-Montenegro D, Yáñez-Sedeño P, Pingarrón JM (2005) Anal

Bioanal Chem 382:381-387. 39. De Sanoit J, Vanhove E, Mailley P, Bergonzo P (2009) Electrochim Acta

54:5688-5693. 40. Tan SM, Chua CK, Pumera M (2013) Analyst 138:1700-1704.


227

Screen-printed electrode based electrochemical detector coupled with ionic liquid dispersive liquid-liquid

microextraction and microvolume back-extraction for determination of mercury in water samples

Elena Fernándeza, Lorena Vidala, Daniel Martín-Yergab, Maria del Carmen Blancob,

Antonio Canalsa and Agustín Costa-Garcíab

a Departamento de Química Analítica, Nutrición y Bromatología e Instituto Universitario de Materiales, Universidad de Alicante, P.O. Box 99, 03080 Alicante, Spain. b Departamento de Química Física y Analítica, Universidad de Oviedo, C/Julián Clavería ,8, 33006 Oviedo, Spain.

Received: 31 July 2014 / Received in revised form: 31 October 2014 / Accepted: 4 November 2014 / Available online: 18 December 2014

Abstract

A novel approach is presented, whereby gold nanostructured screen-printed carbon electrodes (SPCnAuEs) are combined with in-situ ionic liquid formation dispersive liquid-liquid microextraction (in-situ IL-DLLME) and microvolume back-extraction for the determination of mercury in water samples. In-situ IL-DLLME is based on a simple metathesis reaction between a water-miscible IL and a salt to form a water-immiscible IL into sample solution. Mercury complex with ammonium pyrrolidine dithiocarbamate is extracted from sample solution into the water-immiscible IL formed in-situ. Then, an ultrasound-assisted procedure is employed to back-extract the

Talanta 135 (2015) 34-40

Capítulo 4

228

mercury into 10 µL of a 4 M HCl aqueous solution, which is finally analyzed using SPCnAuEs.

Sample preparation methodology was optimized using a multivariate optimization strategy. Under optimized conditions, a linear range between 0.5 and 10 µg L-1 was obtained with a correlation coefficient of 0.997 for six calibration points. The limit of detection obtained was 0.2 µg L-1, which is lower than the threshold value established by the Environmental Protection Agency and European Union (i.e., 2 µg L-1 and 1 µg L-1, respectively). The repeatability of the proposed method was evaluated at two different spiking levels (3 and 10 µg L-1) and a coefficient of variation of 13% was obtained in both cases. The performance of the proposed methodology was evaluated in real-world water samples including tap water, bottled water, river water and industrial wastewater. Relative recoveries between 95% and 108% were obtained.

Keywords: Liquid-phase microextraction, dispersive liquid-liquid microextraction, ionic liquid, mercury, screen-printed electrode, water samples.

1. Introduction

Mercury is one of the most well-known toxic elements and even the World Health Organization places it between the first ten chemicals or group of chemicals of major public health concern [1]. Mercury exists in different forms with different properties, namely elemental or metallic (i.e., Hg0); inorganic (i.e., Hg2+); and organic (i.e., MeHg+, EtHg+, PhHg+). Several factors determine the adverse effects from mercury exposure including its chemical form, the dose, the age and health of the person exposed, and the duration and kind of exposure (e.g., inhalation, ingestion, etc.) [2]. Among the most relevant health effects we can mention damage to the gastrointestinal tract, nervous system, kidneys, respiratory failures and problems during the development of organs in unborn.

Mercury enters in the environment through both biogenic and anthropogenic vias. However, human activities such as mining, burning of fossil fuels, agriculture, paper and electrochemical industries, and household


229

wastes, are the main responsible of the concerning increase of mercury levels in air, soil and water of certain contaminated areas. Monitoring the presence of mercury in natural and drinking waters is of great interest due to its high toxicity and bioaccumulation factor [3]. Mercury concentrations are commonly in the range of low ng L-1 in environmental waters [3] whereas the permitted level of mercury in drinking water depends on the responsible authorities of each territory. For example, the Environmental Protection Agency (EPA) sets the threshold level at 2 µg L-1 [4], but the European Union establishes the limit at 1 µg L-1 [5].

Electrochemical techniques have been widely employed to determine mercury in natural and drinking waters. Two excellent reviews have been recently published about the latest advances in electrochemical, mainly voltammetric, determination of mercury [6,7]. Electrochemistry offers sensitivity, simplicity, rapid response and inexpensive instrumentation with miniaturization and portable options. A major drawback to be considered results from the difficulty of removing mercury from electrode surface between measurements which leads to memory effect problems [6,7]. However, tedious and time consuming cleaning steps can be avoided with the use of screen-printed electrodes (SPELs), which can be disposable after a single use due to their high cost effectiveness. Several methods based on SPELs have been reported for the determination of mercury in different water samples, including the use of bare gold SPELs [8], and modified SPELs with carbon nanomaterials [9-11], gold films [12,13], gold nanoparticles [14,15], nanohybrid materials [14] and chelating agents [16]. As can be seen in Table 1, the vast majority of the reported works include a preconcentration step over the working electrode (i.e., deposition time) followed by anodic stripping voltammetry. Gold is commonly employed in working electrodes due to its high affinity for mercury which leads to an improvement in its preconcentration. In addition, mercury suffers from a process named underpotential deposition (UPD) on gold electrodes [7]. The presence of gold promotes the adsorption of mercury atoms on the surface once the ionic metal is reduced forming an amalgam (Au-Hg). The formation of this amalgam is energetically more favored with respect to pure mercury and makes the deposition of mercury on gold occur at a more positive potential than in normal

Capítulo 4

230

conditions. As a consequence, the selectivity of the method is generally improved. In this work, screen-printed carbon electrodes modified with gold nanoparticles (SPCnAuEs) are employed as electrochemical transducers in the detection stage. The use of nanoparticles in electroanalysis is continuously growing due to its numerous advantages, related to the unique properties of nanoparticulate materials [17] (e.g., increasing surface area, enhanced mass transport and improving selectivity, catalytic activity and signal to noise ratio).

Liquid-phase microextraction (LPME) [18] appeared in the latest nineties offering undoubted advantages as miniaturized extraction techniques, such as simplicity, easiness to handle, low sample and solvent consumptions, and an important reduction of residues generated. One of the most popular LPME technique is dispersive liquid-liquid microextraction (DLLME) [19] which has even come to dominate LPME research publications in the recent years [20]. DLLME is based on the complete dispersion of the small volume of extractant solvent into the sample, normally assisted by a disperser agent. During DLLME, there is a high contact between phases therefore the extraction is really rapid and effective. After the extraction, phases are separated normally by centrifugation and the enriched phase with analyte is analyzed. Numerous modifications of the original DLLME procedure [19] have been reported up to now [21] including the use of new extractant solvents such as ionic liquids (ILs) [22]. Within the use of ILs, a novel methodology called in-situ IL formation dispersive liquid-liquid microextraction (in-situ IL-DLLME) [23,24] has recently been developed. In-situ IL-DLLME is based on the formation of a water-immiscible IL using a metathesis reaction between a water-miscible IL and an ion exchange salt into sample solution. Thereby, the extractant phase is generated in-situ in form of homogeneously dispersed fine drops, the disperser agent is totally avoided and the extraction efficiency generally increases.

Different LPME techniques including single-drop microextraction [25,26], DLLME [27-29], in-situ IL-DLLME [23] and task-specific IL ultrasound-assisted DLLME [30] have been employed for the determination and speciation of mercury in water samples. In these works, bulky and expensive chromatographic systems [25,28,29], capillary electrophoresis [27], UV-vis spectrometry [23], cold vapor [30] and electrothermal vaporization atomic


231

absorption spectrometry [26] were used as separation and detection techniques, respectively.

The approach presented here employs an in-situ IL-DLLME followed by an ultrasound-assisted microvolume back-extraction and SPCnAuEs as electrochemical transducers for the determination of mercury in water samples. This combination exploits the advantages of including a miniaturized sample preparation step with the high sensitivity and specificity that offers the electrochemical determination of mercury using SPCnAuEs. LPME provides a high preconcentration of the analyte and a clean-up step for dirty matrices employing low amounts of sample and chemicals. In addition, considering the low volume of sample needed for analysis with SPELs, they appear as an alternative and perfectly compatible detection methodology after miniaturized extraction techniques, thus avoiding classical and bulky analytical instrumentation [31]. A multivariate optimization strategy has been adopted for the optimization of the sample preparation and the applicability of the method has been tested studying real-world water samples.

Capítulo 4

232

Table 1. Comparison of different methods using SPELs for the determination of mercury in water samples.

Electrode Lineal range LOD Real water samples

Comments/Analytical technique (deposition time in parentheses) Ref.

SPGE 5-30 ng mL-1 1.1 ng mL-1 Wastewater and rain water SWASV (60 s) [8]

SPEL/carbon black 2.5·10-8-1·10-7 M (5-20 µg L-1)

5·10-9 M (1 µg L-1) Drinking water Indirect determination by amperometric

measurements of thiols [9]

SPBE/MWCNTs 0.2-40 µg L-1 0.09 µg L-1 Tap water SWASV (180 s) [10]

Carbon NPs-based SPELs 1-10 µg L-1 - Seawater Heated electrodes/ SWASV (120 s) [11]

SPEL/gold film 2-16 µg L-1 1.5 µg L-1 Tap water SWASV (120 s) [12]

SPEL/gold film 0.2-0.8 µg L-1 0.08 µg L-1 - Preconcentration step using magnetic nanoparticles modified with thiols/SWASV (120 s) [12]

SPCE/gold film 0-100 µg L-1 0.9 µg L-1 - SWASV (120 s) [13]

SPGOnAuEs 2-50 µg L-1 1.9 µg L-1 - SWASV (200 s) [14]

SPGE, screen-printed gold electrode; SWASV, square-wave anodic stripping voltammetry; SPBE, screen-printed bismuth electrode; MWCNTs, multi-walled carbon nanotubes; NPs, nanoparticles; SPGOnAuEs, screen-printed graphene oxide/gold nanoparticles electrodes.


233

Table 1 (cont.). Comparison of different methods using SPELs for the determination of mercury in water samples.

Electrode Lineal range LOD Real water samples Comments/Analytical technique (deposition time in parentheses) Ref.

SPCNTnAuEs 0.5-50 µg L-1 0.2 µg L-1 Tap and river waters SWASV (200 s) [14]

SPCnAuEs 5-100 µg L-1 3.3 µg L-1 - SWASV (240 s) [14]

SPCnAuEs 5-20 ng mL-1 0.8 ng mL-1 Rain and river

waters, industrial wastewater

SWASV (120 s) [15]

CTS-SPEL 20-80 ng mL-1 2 ng mL-1 - DPASV (30 s) [16]

SPCNTnAuEs, screen-printed carbon nanotubes/gold nanoparticles electrodes; SWASV, square-wave anodic stripping voltammetry; CTS-SPEL, chitosan-modified screen-printed electrodes; DPASV, differential pulse anodic stripping voltammetry.

Capítulo 4

234


2.1. Reagents and water samples

A stock standard solution of Hg2+ (1000 mg L-1) was prepared by dissolving Hg(OAc)2 (≥ 99%) from Fluka (Stenhein, Germany) in ultrapure water. Working solutions were prepared by proper dilution of the stock standard. The 1-hexyl-3-methylimidazolium chloride IL ([Hmim][Cl]) (98%) was purchased from Iolitec (Heilbronn, Germany). The lithium bis[(trifluoromethyl)sulfonyl]imide (LiNTf2) salt and the chelating agent ammonium pyrrolidine dithiocarbamate (APDC) (~99%) were supplied by Sigma-Aldrich (St. Louis, MO, USA). A solution of 2 mg mL-1 of the chelating agent was prepared by dissolving APDC in ultrapure water. Reactive grade NaCl and NaOH (≥97%, pellets) were from ACS Scharlau (Barcelona, Spain). Fuming HCl (37%) was supplied by Merck (Madrid, Spain). The ultrapure water employed for preparing all solutions was obtained with a Millipore Direct System Q5™ purification system from Ibérica S.A. (Madrid, Spain).

Standard Au3+ tetrachloro complex (1.000±0.002 g of AuCl4- in 500 mL of

1.0 M HCl) was purchased from Merck. Solutions of 1 mM AuCl4- were

prepared by suitable dilution of this standard solution in 0.1 M HCl. Tap water was collected from the water-supplied network of the lab in the

Department of Physical and Analytical Chemistry of the University of Oviedo (Spain). Bottled water (San Benedetto mineral water, Valencia, Spain) was purchased in the supermarket. River water from Nora river was collected in Tiñana (Siero, Spain) and industrial wastewater was from Galicia (Spain). The wastewater contained a chemical oxygen demand of 7 mg O2 L-1 and ˂5 mg L-1 of suspended solids. All water samples were stored at 4 ºC and were used without any further pretreatment. Initial analysis with the developed method confirmed that mercury levels were undetectable in the four selected water samples.

2.2. Apparatus and electrodes

An ultrasounds bath from Elma (Singen, Germany) was used to assist the back-extraction procedure.


235

An Autolab PGSTAT 12 potentiostat from Eco Chemie (Ultrecht, The Netherlands) controlled by Autolab GPES software version 4.8 was used for electrochemical experiments.

Screen-printed carbon electrodes (SPCEs) (ref. DRP-110) with three electrode configuration were purchased from DropSens (Oviedo, Spain). The working electrode, with a disk-shaped of 4 mm of diameter, and the counter electrode were made of a carbon ink whereas the pseudo-reference electrode was made of silver. Specific connectors obtained from DropSens (ref. DRP-DSC) were used for the conexion of the SPCEs to the potentiostat.

2.3. In-situ IL-DLLME and microvolume back-extraction

Under optimum conditions, 20 mg of [Hmim][Cl] were placed in a test tube and dissolved in 4 mL of aqueous standards or sample solutions and the chelating agent (40 µL of 2 mg mL-1). The ionic exchange salt LiNTf2 was added in an equimolar ratio (i.e., 28.3 mg) with the IL [Hmim][Cl], according to previous works [24,31]. A cloudy solution was immediately formed and the mixture was manually shaken for 0.5 min. In order to accelerate phases separation, the tube was then introduced in an ice bath for 5 min. Next, the phases were separated by centrifugation for 10 min at 5000 rpm. The aqueous phase was removed with a glass pipette, and the formed IL-phase (i.e., 20 µL of [Hmim][NTf2]) was withdrawn with a micropipette and deposited in an Eppendorf tube of 0.5 mL. For the back-extraction, 10 µL of 4 M HCl aqueous solution was added to the IL phase and the mixture was sonicated in an ultrasounds bath for 14 min at 90% of power and 37 KHz of frequency. Since direct measurements on the IL were not suitable, back-extraction was necessary for voltammetric analysis. After back-extraction, phases were separated by centrifugation for 5 min at 5000 rpm and the enriched acidic aqueous phase that remained in the upper part was analyzed. The overall procedure is graphically described in Figure 1.

Capítulo 4

236

Figure 1. In-situ IL-DLLME and ultrasound-assisted microvolume back-extraction coupled with SPCnAuE.


Gold nanoparticles were generated over the SPCEs surface employing the procedure developed by Martínez-Paredes et al. [32] and previously optimized by Martín-Yerga et al. for the determination of mercury [14]. Briefly, 40 µL of a 1 mM AuCl4- solution in 0.1 M HCl aqueous solution were dropped onto the SPCE surface and a constant current of -100 µA was applied for 180 s. After

[Hmim][NTf2] Sample addition (Hg and APDC)

LiNTf2 addition

Ice bath

Shaking

[Hmim][Cl]

Centrifugation

Removing aqueous phase

Addition of HCl Sonication

Centrifugation

Withdrawal the upper enriched acid phase

Voltammetric analysis with SPCnAuEs

Transfer to Eppendorf tube


237

gold nanoparticles deposition, the electrode surface was generously rinsed with ultrapure water and dried at room temperature before use. A new SPCnAuEs was prepared and employed for each experiment. The electrochemical behavior of mercury on SPCnAuEs was previously and deeply studied [14], therefore, no further discussion will be included in the present work. After back-extraction, 5 µL of the resulting upper acidic aqueous phase was mixed with 37 µL of 0.5 M NaOH in order to obtain a suitable electrolytic medium. A volume of 40 µL of this solution was deposited on the electrode surface for voltammetric measurements. Mercury was determined by square-wave anodic stripping voltammetry employing previous optimized conditions [14]. Mercury was preconcentrated over SPCnAuEs by applying a constant potential of +0.3 V for 240 s. Thereafter, the potential was recorded between +0.3 V and +0.55 V at a frequency of 80 Hz, amplitude of 30 mV and step potential of 4 mV. All experiments were carried out at room temperature.


An anodic peak corresponding to the reoxidation of mercury appears at approximately +0.42 V and the height of this peak was employed for the quantification of the analyte. The “base line correction” option provided by GPEs software was employed to get more defined peaks, specially at low concentrations, and to obtain more reliable and accurate measurements. A two-step multivariate optimization strategy, using Plackett-Burman and central composite designs, was carried out to determine the optimum conditions of sample preparation. Minitab 15 statistical software (State College, PA, USA) was employed to construct the experimental design matrices and evaluate the results.

Capítulo 4

238


3.1. Optimization of sample preparation

3.1.1. Screening step

Plackett-Burman design is a two-level fractional factorial design that ignores interaction between factors and therefore main effects can be calculated with a reduced number of experiments leading to a saving in resources and time. The Plackett-Burman design results are very useful in the first steps of a project when many factors are initially considered but finally only a few show important effects [33]. A saturated Plackett-Burman design was used to construct the matrix of experiments, including 11 factors: eight real factors and three dummy factors. The effects of dummy factors were used to evaluate the experimental error [34,35]. The eight real experimental factors selected at two levels were: amount of [Hmim][Cl], amount of chelating agent, ionic strength, sample pH, volume of HCl acceptor solution during back-extraction, back-extraction time, power and frequency of the ultrasounds bath. Table 2 shows the experimental factors and levels considered in the Plackett-Burman design. A total of twelve experiments were randomly performed using aqueous standards of 25 µg L-1.

Table 2. Experimental factors and levels of the Plackettt-Burman design.

Factors Level

Low (-1) High (+1) Amount of [Hmim][Cl] (mg) 20 40

Amount of chelating agent (µL, 2 mg mL-1) 20 40

Ionic strength (NaCl concentration, %, w/v) 0 10

Sample pH 5 10

HCl volume (µL) 20 50

Back-extraction time (min) 5 10

Ultrasounds power (%) 50 90

Ultrasounds frequency (KHz) 37 80


239

The data obtained were evaluated using an ANOVA test and the results were visualized with the Pareto chart shown in Figure S1. The length of each bar was proportional to the influence of the corresponding factor and the effects that exceed each reference vertical line can be considered significant with 95% and 90% probability, respectively.

Figure S1. Pareto chart of the Plackett-Burman design.

According to Figure S1, the ultrasounds frequency and HCl volume were statistically significant factors, with 95% probability, showing a negative effect. The negative effect of the frequency is in agreement with the fact that at high ultrasounds frequencies, cavitation bubbles are more difficult to create as a result of the shorter duration of rarefraction cycles. Higher amplitudes (i.e., power) would be necessary to ensure that cohesive forces in the liquid were overcome and maintain a certain cavitational energy [36]. For the HCl volume, the negative effect is easily explained considering that if less volume of acid is used, a higher concentration of the analyte is obtained in the final acceptor solution. The ultrasounds device employed during this work only accepted two discrete values of frequency, namely 37 and 80 KHz, thus this significant factor could not be included in the following optimization step and was fixed in

0 2 4 6 8 10 12 14 16

Dummy 3Ionic strength (% NaCl)

Dummy 1Dummy 2

Ultrasonic power (%)pH

Chelating agent (µL)Amount of [Hmim][Cl] (mg)

Back-extraction time (min)HCl volume (µL)

Ultrasonic frequency (kHz)

Effects

Possitive effect Negative effect

95% 90%

Capítulo 4

240

its lower level. As a consequence, back-extraction time and amount of [Hmim][Cl], which showed significant effects with 90% probability (see Figure S1), were included in the next optimization step. Back-extraction time showed a positive effect whereas for the amount of [Hmim][Cl] the effect was negative. These effects revealed that the mass transfer during back-extraction is not as instantaneous as during in-situ IL-DLLME, and it is enhanced if low amounts of [Hmim][NTf2] are formed, probably related to diffusion effects from the bulk of the IL to the contact surface with the HCl aqueous solution.

The other four real factors considered in screening step with non-significant effects were fixed at the following levels: amount of chelating agent, 40 µL (2 mg mL-1); ionic strength, no addition of NaCl; sample pH, the pH of water without any adjustment; and ultrasounds power, 90%.

3.1.2. Optimization of significant factors

Central composite design (CCD) combines a two-level full factorial design (2k) with 2k star points, where k is the number of factors being optimized, and one point at the center of the experimental region. In order to ensure the

rotatability of the model, star points were set at α=√2k4=1.682 whereas the

central point was repeated five times to provide an orthogonal design [33]. CCD was used to evaluate and optimize main effects, interaction effects and quadratic effects of the three considered factors. Table 3 shows the low and high levels, the central and star points of the considered factors in the optimization step. Nineteen experiments were randomly performed using aqueous standards of 25 µg L-1.


241

Table 3. Factors, low and high levels, central and star points used in CCD design.

The data obtained were also evaluated using an ANOVA test. The

coefficients of the factors and the p-values are listed in Table S1.

Table S1. Coefficients and p-values obtained in CCD design.

Factors Coefficient p-value HCl volume (C1) -15.429 0*

Back-extraction time (C2) 8.581 0.002* Amount of [Hmim][Cl] (C3) -3.668 0.101

C1*C1 -0.572 0.869 C2*C2 -13.322 0.003* C3*C3 -11.672 0.007* C1*C2 -2.863 0.533 C1*C3 -3.429 0.458 C2*C3 2.722 0.553

(*) Significant factors with 95% probability (i.e., p-value < 0.05).

Significant factors with 95% probability (i.e., p-value ˂ 0.05) were HCl volume, back-extraction time and the quadratic effects of back-extraction time and amount of [Hmim][Cl], which confirms the curvature of the system and its fitting with the proposed second-grade polynomial system. The adjustment obtained expressed as r2 value was 92%.

Factor Level

Star points (1.682)

Low (-1)

Central (0)

High (+1) - +

HCl volume (µL) 22 40 58 10 70

Back-extraction time (min) 6 10 14 3 17

Amount of [Hmim][Cl] (mg) 28 40 52 20 60

Capítulo 4

242

The response surfaces obtained using the CCD are shown in Figure 2. Pairs of factors were considered separately in order to easily interpret the effect of each one on the response of the system. Thus, Figure 2a shows the response surface which results in plotting HCl volume vs. back-extraction time, for 40 mg of [Hmim][Cl]; Figure 2b shows the response surface obtained as a function of HCl volume and amount of [Hmim][Cl], whilst back-extraction time is fixed at 10 min; and Figure 2c shows the surface response corresponding the effects of back-extraction time and amount of [Hmim][Cl], with established HCl volume at 40 µL. As expected, HCl volume has a negative effect (Figure 2a and b) and the response of the system increases when the HCl volume decreases. For the back-extraction time, the response of the system increases with the time (Figure 2a and c) until reaching a maximum at 14 min. Both, 10 µL for HCl volume and 14 min for the back-extraction time, were adopted as the optimum conditions for the proposed methodology. As can be seen in Figure 2b and c, the effect of the amount of [Hmim][Cl] also presents a maximum over 40 mg, although the variation of the response is really slight between 40 and 20 mg. Thus, considering the sign of the effect of this factor obtained in the Plackett-Burman design, which was negative, and the importance of waste reduction, 20 mg of [Hmim][Cl] were finally chosen for the validation of the method.

In summary, the results obtained from the overall optimization process lead to the following experimental conditions: amount of [Hmim][Cl], 20 mg; amount of chelating agent, 40 µL (2 mg mL-1); ionic strength, no addition of NaCl; sample pH, the pH of water without any adjustment; HCl volume, 10 µL; back-extraction time, 14 min; ultrasounds power, 90%; and ultrasounds frequency, 37 KHz.


243

(a)

(b)

Figure 2. Surfaces response of CCD design obtained by plotting: (a) HCl volume vs. back-extraction time (amount of [Hmim][Cl]: 40 mg); (b) HCl volume vs. amount of [Hmim][Cl] (back-extraction time: 10 min); (c) amount of [Hmim][Cl] vs. back-extraction time (HCl volume: 40 µL).

Peak

hei

ght (

µA)

HCl volume (µL)

Back-extraction time (min) 5

10 15

20 40 60

20

40

60

60 40

20 70 40

10

20

40

60

Peak

hei

ght (

µA)

HCl volume (µL)

Amount of [Hmim][Cl] (mg)

Capítulo 4

244

(c)

Figure 2 (cont). Surfaces response of CCD design obtained by plotting: (a) HCl volume vs. back-extraction time (amount of [Hmim][Cl]: 40 mg); (b) HCl volume vs. amount of [Hmim][Cl] (back-extraction time: 10 min); (c) amount of [Hmim][Cl] vs. back-extraction time (HCl volume: 40 µL).


Quality parameters of the proposed method were evaluated. Under optimized conditions, a concentration range from 0.5 to 25 µg L-1 was studied. Finally, the linear working range was established between 0.5 and 10 µg L-1. The calibration curve was constructed using six concentration levels, evaluated by triplicate. The voltammograms corresponding to the blank and the aqueous standards of concentrations from 0.5 to 10 µg L-1 are shown in Figure 3.

Peak

hei

ght (

µA)

10

20

30

40

20 40

60 5

10

15

Amount of [Hmim][Cl] (mg)

Back-extraction time (min)


245

Figure 3. Square-wave voltammograms, after baseline correction, of a blank and mercury aqueous standards of 0.5, 1, 3, 5, 7 and 10 µg L-1 after in-situ IL-DLLME and back-extraction under optimum conditions.

The resulting calibration curve gave a high level of linearity with a

correlation coefficient (r) of 0.997 (N=6). The sensitivity of the instrumental measurements estimated by the slope of the calibration curve was (3.0±0.3) µA µg-1 L. The repeatability of the proposed method, expressed as coefficient of variation (CV), was evaluated by five consecutive analyses of aqueous standards at concentrations of 3 and 10 µg L-1. CV values of 13% were found in both cases. An enrichment factor of 25 was obtained for the proposed procedure, defined as the slope ratio of the calibration curves with and without preconcentration. The limit of detection (LOD) was estimated according to the Directive 98/83/EC [5], on the quality of water intended for human consumption, as the concentration corresponding to a signal that is five times the standard deviation of the blank. The LOD was found to be 0.2 µg L-1, which is lower than most of the reported works up to now using SPELs (see Table 1), and stands lower than the threshold value established by both, the

0.32 0.36 0.40 0.44 0.48 0.520

5

10

15

20

25

30

35

I (

A)

E (V)

10 g L-1

7 g L-1

5 g L-1

3 g L-1

1 g L-1

0.5 g L-1

blank

Capítulo 4

246

EPA and the European Union (i.e., 2 µg L-1 and 1 µg L-1, respectively). It is important to point out that the sensitivity and LOD of the proposed method are significantly better than those obtained in a previous work [14] (i. e., 0.120 µA µg-1 L and 3.3 µg L-1, respectively) using the same kind of SPCnAuEs under equal conditions but without sample preparation. In addition, just a few of the reported works using SPEs have shown equal or lower LOD than 0.2 µg L1 [10,12,14], however, standard addition method was needed to analyze real-world water samples [12,14]. Therefore, the great but scarcely explored advantages that offer the combination of LPME with electrochemical detection using SPELs have been demonstrated.

3.3. Real-world water samples analysis

The applicability of the proposed method to determine mercury in real-world water samples was evaluated studying matrix effects. Four water samples (namely tap water, bottled water, river water and wastewater) were employed for recovery studies. As mentioned before, previous analysis revealed that mercury levels in the samples were under the LOD of the present approach. Three replicated analysis of each water sample were carried out at two different spiking levels (1 and 7 µg L-1). Relative recoveries were calculated as the ratio of the signals found in real and ultrapure water samples spiked at the same concentration level. As can be observed in Table 4, relative recoveries ranged from 95 to 108% in the four water samples, whereas the CV values were between 7 and 15%. According to these results, it can be concluded that the matrix effects were not significant for the determination of mercury in the four selected water samples.


247

Table 4. Relative recoveries and CV values (in parentheses) for the analysis of mercury in real-world water samples.

Water sample Relative recoveries

1 µg L-1 7 µg L-1 Tap water 106 (11) 108 (7)

Bottled water 98 (11) 103 (15)

River water 97 (10) 98 (9)

Wastewater 97 (12) 95 (9)

4. Conclusions

SPCnAuEs have been successfully combined with in-situ IL-DLLME and microvolume back-extraction methodologies for the determination of mercury in water samples, reaching a LOD that satisfies the established legal threshold levels and proving its applicability in real-world water sample analysis. Higher sensitivity and lower LOD were obtained with the proposed methodology compared to those obtained with the same electrochemical transducers but omitting the sample preparation. Therefore, the great and up to now practically unexplored benefits that offer the combination of miniaturized sample preparation techniques with the electrochemical analysis using SPELs have been experimentally demonstrated. Although the ice bath, centrifugation and sonication limit the in-field application of the proposed methodology, authors strongly believe in a promising future for the synergistic combination of LPME with SPELs as detection methodology within the perspectives of developing inexpensive analytical methodologies with portable options for rapid and on-site measurements.

Acknowledgments

The authors would like to thank the Spanish Ministry of Science and Innovation (Project nos. CTQ2011-23968 and CTQ2011-24560), the Generalitat Valenciana (Spain) (Project nos. ACOMP/ 2013/072,

Capítulo 4

248

PROMETEO/2013/038 and GV/2014/096) and the University of Alicante (Spain) (Project no. GRE12-45) for the financial support. E. Fernández and D. Martín-Yerga also thank Generalitat Valenciana and Ministry of Economy and Competitiveness, respectively, for their fellowships.

References

[1] World Health Organization, June 2014, ⟨http://www.who.int/ipcs/assessment/public_health/chemicals_phc/en/index.html⟩.

[2] World Health Organization, June 2014, ⟨http://www.who.int/mediacentre/ factsheets/fs361/en/⟩.

[3] K. Leopold, M. Foulkes, P.J. Worsfold, TrAC Trends Anal. Chem. 28 (2009) 426-435.

[4] Environmental Protection Agency, June 2014, ⟨http://water.epa.gov/drink/contaminants/basicinformation/mercury.cfm⟩.

[5] Council Directive 98/83/EC of 3 November on the quality of water intended for human consumption.

[6] C. Gao, X.-J. Huang, TrAC Trends Anal. Chem. 51 (2013) 1-12. [7] D. Martín-Yerga, M.B. González-García, A. Costa-García, Talanta 116

(2013) 1091-1104. [8] E. Bernalte, C. Marín Sánchez, E. Pinilla Gil, Anal. Chim. Acta 689

(2011) 60-64. [9] F. Arduini, C. Majorani, A. Amine, D. Moscone, G. Palleschi,

Electrochim. Acta 56 (2011) 4209-4215. [10] X. Niu, H. Zhao, M. Lan, Anal. Sci. 27 (2011) 1237-1241. [11] G. Aragay, J. Pons, A. Merkoçi, J. Mater. Chem. 21 (2011) 4326-4331. [12] A. Mandil, L. Idrissi, A. Amine, Microchim. Acta 170 (2010) 299-305. [13] S. Laschi, I. Palchetti, M. Mascini, Sens. Actuators B 114 (2006) 460-

465. [14] D. Martín-Yerga, M.B. González-García, A. Costa-García, Sens.

Actuators B 165 (2012) 143-150. [15] E. Bernalte, C.M. Sánchez, E.P. Gil, Sens. Actuators B 161 (2012) 669-

674.


249

[16] E. Khaled, H.N.A. Hassan, I.H.I. Habib, R. Metelka, Int. J. Electrochem. Sci. 5 (2010) 158-167.

[17] F.W. Campbell, R.G. Compton, Anal. Bioanal. Chem. 396 (2010) 241-259.

[18] F. Pena-Pereira, I. Lavilla, C. Bendicho, TrAC Trends Anal. Chem. 29 (2010) 617-628.

[19] M. Rezaee, Y. Assadi, M.-R. Milani Hosseini, E. Aghaee, F. Ahmadi, S. Berijani, J. Chromatogr. A 1116 (2006) 1-9.

[20] J.M. Kokosa, TrAC Trends Anal. Chem. 43 (2013) 2-13. [21] H. Yan, H. Wang, J. Chromatogr. A 1295 (2013) 1-15. [22] M.J. Trujillo-Rodríguez, P. Rocío-Bautista, V. Pino, A.M. Afonso, TrAC

Trends Anal. Chem. 51 (2013) 87-106. [23] M. Baghdadi, F. Shemirani, Anal. Chim. Acta 634 (2009) 186-191. [24] C. Yao, J.L. Anderson, Anal. Bioanal. Chem. 395 (2009) 1491-1502. [25] F. Pena-Pereira, I. Lavilla, C. Bendicho, L. Vidal, A. Canals, Talanta 78

(2009) 537-541. [26] H. Bagheri, M. Naderi, J. Hazard. Mater. 165 (2009) 353-358. [27] J. Li, W. Lu, J. Ma, L. Chen, Microchim. Acta 175 (2011) 301-308. [28] X. Jia, Y. Han, X. Liu, T. Duan, H. Chen, Spectrochim. Acta Part B 66

(2011) 88-92. [29] Z. Gao, X. Ma, Anal. Chim. Acta 702 (2011) 50-55. [30] E. Stanisz, J. Werner, H. Matusiewicz, Microchem. J. 110 (2013) 28-35. [31] E. Fernández, L. Vidal, J. Iniesta, J.P. Metters, C.E. Banks, A. Canals,

Anal. Bioanal. Chem. 406 (2014) 2197-2204. [32] G. Martínez-Paredes, M.B. González-García, A. Costa-García,

Electrochim. Acta 54 (2009) 4801-4808. [33] D.C. Montgomery, Design and Analysis of Experiments, seventh ed.,

Wiley, Hoboken, New Jersey, United States, 2009. [34] Y.V. Heyden, C. Hartmann, D.L. Massart, L. Michel, P. Kiechle, F. Erni,

Anal. Chim. Acta 316 (1995) 15-26. [35] H. Fabre, N. Mesplet, J. Chromatogr. A 897 (2000) 329-338. [36] M.D. Priego Capote, F. Luque de Castro, Analytical Applications of

Ultrasounds, first ed., Elsevier B.V., Amsterdam, The Netherlands, 2007.


251

Mercury determination in urine samples by gold-

nanostructured screen-printed carbon electrodes after vortex-assisted ionic liquid dispersive liquid-liquid microextraction

Elena Fernándeza, Lorena Vidala, Agustín Costa-Garcíab and Antonio Canalsa

a Departamento de Química Analítica, Nutrición y Bromatología e Instituto Universitario de Materiales, Universidad de Alicante, P.O. Box 99, 03080 Alicante, Spain. b Departamento de Química Física y Analítica, Universidad de Oviedo, C/Julián Clavería ,8, 33006 Oviedo, Spain.

Received: 16 December 2015 / Received in revised form: 17 February 2016 / Accepted: 22 February 2016 / Available online: 23 February 2016

Abstract

A novel approach is presented to determine mercury in urine samples, employing vortex-assisted ionic liquid dispersive liquid-liquid microextraction and microvolume back-extraction to prepare samples, and screen-printed electrodes modified with gold nanoparticles for voltammetric analysis. Mercury was extracted directly from non-digested urine samples in a water-immiscible ionic liquid, being back-extracted into an acidic aqueous solution. Subsequently, it was determined using gold nanoparticle-modified screen-printed electrodes. Under optimized microextraction conditions, standard addition calibration was applied to urine samples containing 5, 10 and 15 µg L-

1 of mercury. Standard addition calibration curves using standards between 0 and 20 µg L-1 gave a high level of linearity with correlation coefficients ranging from 0.990 to 0.999 (N=5). The limit of detection was empirical and

Analytica Chimica Acta 915 (2016) 49-55

Capítulo 4

252

statistically evaluated, obtaining values that ranged from 0.5 to 1.5 µg L-1, and from 1.1 to 1.3 µg L-1, respectively, which are significantly lower than the threshold level established by the World Health Organization for normal mercury content in urine (i.e., 10-20 µg L-1). A certified reference material (REC-8848/Level II) was analyzed to assess method accuracy finding 87% and 3 µg L-1 as the recovery (trueness) and standard deviation values, respectively. Finally, the method was used to analyze spiked urine samples, obtaining good agreement between spiked and found concentrations (recovery ranged from 97 to 100%).

Keywords: vortex-assisted dispersive liquid-liquid microextraction, ionic liquid, mercury, screen-printed electrodes, urine samples.

1. Introduction

Mercury is a highly toxic element whose adverse health effects depend on several factors such as chemical form, route of exposure, dose and personal features [1]. Inhalation exposure mainly corresponds to elemental mercury (i.e., Hg0) due to its high vapor pressure. Occupational exposure to Hg0 vapors occurs in mining and fossil-fuel processing activities, manufacture of amalgams, manipulation of mercury-containing fungicides, waste incineration or chloralkali plants. Hg0 is oxidized to Hg2+ in most body tissues and can be retained and accumulated, especially in the brain and kidneys. Oral intake is the main source of inorganic mercury (i.e., Hg2+), although its absorption from gastrointestinal tract occurs only to a limited extent [2]. Cutaneous absorption has been proposed as another less significant route of exposure, since dermal penetration of Hg2+ can occur through use of skin-lightening cosmetic products containing mercuric salts. Once in the body, Hg2+ accumulates mainly in the kidneys. Methylmercury (i.e., MeHg+) is the most toxic and frequent form of organic mercury. MeHg+ exposure mainly occurs through a diet high in fish and marine mammals. In contrast to Hg2+, MeHg+ is rapidly and extensively absorbed through the gastrointestinal tract and accumulates predominantly in the brain [2].


253

Urine and blood have been broadly employed for risk assessment of mercury exposure and health risk prevention. Mercury content in urine generally reflects recent exposure to inorganic and/or elemental mercury. However, Hg2+ accumulates in the kidneys and is slowly excreted through urine, therefore, urinary mercury can also reflect long-term exposure in the past [2]. MeHg+ is mostly eliminated by demethylation and excretion in the feces and it is not typically found in urine [1]. Urinary mercury levels are normally expected to be lower than 10-20 µg L-1 in an unexposed population.

Different publications report mercury determination in urine using cold-vapor atomic absorption [3] or fluorescence [4] spectrometry, electrothermal absorption spectrometry [5], UV-Vis spectrophotometry [6], inductively coupled plasma atomic emission [7] or mass [8] spectrometry. Besides spectrometric techniques, electrochemical techniques have also been proposed [9-13]. Electrochemistry offers sensitivity, simplicity, rapid response and inexpensive instrumentation with miniaturization and portable options. In this respect, screen-printed electrodes (SPELs) [14] have gained widespread interest. SPELs are size-reduced devices designed to analyze low-volume samples, which also allow de-centralized testing. In addition, SPELs are mass-produced at a low cost and are thus disposable. In this work, screen-printed carbon electrodes modified with gold nanoparticles (SPCnAuEs) have been employed as electrochemical transducers for mercury determination. Gold nanoparticles exploit the properties of gold as a high affinity material for mercury, with the advantages of including nanosized particles, such as high active surface area, enhanced mass transport and signal to noise ratio [15]. In addition, mercury undergoes a process named underpotential deposition (UPD) on gold electrodes. The presence of gold promotes the adsorption of mercury atoms on the surface once the ionic metal is reduced, forming an amalgam (Au-Hg). This adsorption is usually limited to a monolayer. Due to the strong interaction between gold substrate and reduced mercury, the deposition of mercury is favored energetically and takes place at a less negative potential than the reversible Nernst potential for bulk deposition.

Due to the complexity of biological samples, including urine, sample preparation is necessary prior to electrochemical analysis. To date, the electrochemical methods proposed to determine mercury in urine samples

Capítulo 4

254

employ initial digestion steps to decompose organic matter, which generally involve wet acid digestion [9-13]. However, these procedures constitute a risk for mercury loss and thus careful manipulation is essential to avoid analyte evaporation. In this work, dispersive liquid-liquid microextraction (DLLME) is presented as a valuable alternative for sample preparation. DLLME is a miniaturized liquid-phase extraction technique whose major advantages include: speed and ease of use, low cost, low sample volume, extremely low solvent consumption, reduced generation of wastes, high enrichment factors and affordability. Classical DLLME is based on the dispersion in tiny droplets of a water-immiscible solvent into the aqueous sample with the aid of a disperser agent [16]. Other formats of DLLME are based on vortex agitation [17], ultrasound energy [18], temperature changes [19], metathesis reactions [20] or air-assisted methodology [21]. The cloudy solution formed presents a great contact surface area between the donor and acceptor phases, thus enhancing extraction efficiency. In addition to conventional organic solvents, ionic liquids (ILs) have been employed as extractant phase in DLLME (i.e., IL-DLLME) due to their remarkable properties, such as low vapor pressure, good extractability of organic and inorganic compounds, non-flammability and adjustable hydrophobicity [22].

The purpose of this work is to present a novel method for mercury determination in urine samples, combining vortex-assisted IL-DLLME with electrochemical detection by SPCnAuEs. Mercury complexes with ammonium pyrrolidine dithiocarbamate (APDC) are directly extracted from non-digested urine samples into a water-immiscible IL using vortex agitation. Then, mercury is back-extracted into 10 µL of an acidic aqueous solution, which is finally analyzed by anodic stripping voltammetry. The proposed method is based on a previous work [23], in which mercury was determined in water samples, where some changes related with the microextraction techniques are proposed. In the previous work, mercury was extracted from water samples using an in-situ ionic liquid formation dispersive liquid-liquid microextraction [23]. This microextraction technique was not suitable for urine samples since the formation of a precipitate in the extractant phase formed in-situ hindered its retrieval. Hence, vortex-assisted IL-DLLME was adopted in order to overcome the problem. On the other hand, ultrasound-assisted back-extraction [23] has


255

been replaced by vortex agitation in this work to assist back-extraction of mercury to the final aqueous phase, leading to shorter extraction time.

The present method synergistically combines the advantages of an environmentally friendly miniaturized sample preparation protocol with speed, low cost, high sensitivity and selectivity of the electrochemical detection with SPCnAuEs. To our knowledge, this is the first report of an analytical method in which SPELs are employed to determine mercury in urine samples. The aforementioned method was evaluated in order to demonstrate its applicability to the analysis of real urine samples.


2.1. Reagents and samples

A stock standard solution of Hg2+ (1000 mg L-1 in 2% HNO3) was obtained from High-purity Standards (Charleston, SC, USA). Working solutions were prepared by proper dilution of this stock standard. IL 1-hexyl-3-methylimidazolium bis[(trifluoromethyl)sulfonyl]imide ([Hmim][NTf2]) (99%) was purchased from Iolitec (Heilbronn, Germany). The chelating agent APDC (~99%) was supplied by Sigma-Aldrich (St. Louis, MO, USA). A solution of 2 mg mL-1 of the chelating agent was prepared by dissolving APDC in ultrapure water. Fuming HCl (37%) was supplied by Merck (Madrid, Spain) and used to prepare HCl solution (4 M) in ultrapure water. Reactive grade NaOH (≥97%, pellets) was from ACS Scharlau (Barcelona, Spain) and used to prepare NaOH solution (0.5 M). Reactive grade NaCl was also from ACS Scharlau. The ultrapure water (resistivity of 18.2 MΩ cm at 25 ºC) employed for preparing all solutions was obtained with a Millipore Direct System Q5™ purification system from Ibérica S.A. (Madrid, Spain).

Standard Au3+ tetrachloro complex (1.000±0.002 g of AuCl4- in 500 mL of

1.0 M HCl) was purchased from Merck. Solutions of 1 mM AuCl4- were prepared by suitable dilution of this standard solution in 0.1 M HCl.

Urine samples from healthy human volunteers, unexposed to mercury, were collected in sterilized containers. Urine samples were filtered before use

Capítulo 4

256

and stored at 4 ºC. Preliminary analysis using this method revealed that mercury levels were undetectable in all donated urine samples.

The certified reference material “urine control lyophilized for trace elements” REC-8848/Level II was purchased from LGC Standards (Barcelona, Spain).

2.2. Apparatus and electrodes

A vortex mixer from Heidolph (Schwabach, Germany) was used to assist IL-DLLME and microvolume back-extraction. A centrifuge from Selecta (Barcelona, Spain) was used for phases separation after each extraction.

A Multi Autolab/M101 Potentiostat/Galvanostat from Metrohm Autolab B.V. (Utrecht, The Netherlands) controlled by NOVA software version 1.10 was used for electrochemical experiments. Screen-printed carbon electrodes (SPCEs) (ref. DRP-110) with three-electrode configuration were purchased from DropSens (Oviedo, Spain). The working electrode, with a disk-shaped of 4 mm of diameter, and the counter electrode were made of a carbon ink whereas the pseudo-reference electrode was made of silver. Specific connectors obtained from DropSens (ref. DRP-DSC) were used to connect SPCEs to the potentiostat.

2.3. Vortex-assisted IL-DLLME and microvolume back-extraction

For vortex-assisted IL-DLLME, 10 mL of aqueous standards or urine samples were placed in a test tube and mixed with 40 µL of APDC solution (2 mg mL-1) for analyte complexation. Then, 0.06 g of the water-immiscible IL used as extractant phase (i.e., [Hmim][NTf2]) were added and the mixture was vortexed for 3 min. Next, phases were separated by centrifugation for 10 min at 4000 rpm. The upper aqueous phase was carefully removed with a glass pipette and the IL phase deposited in the bottom of the test tube was transferred to a 0.5 mL Eppendorf tube for microvolume back-extraction. Since direct measurements on the IL were not suitable, back-extraction was necessary for voltammetric analysis [23]. Once the IL was transferred to the Eppendorf tube, 10 µL of 4 M HCl aqueous solution were added and the mixture was shaken


257

using a vortex mixer for 9 min. After centrifugation for 2 min at 4000 rpm, the upper acidic aqueous phase enriched with the analyte was retrieved for final analysis by square-wave anodic stripping voltammetry (SWASV) using SPCnAuEs.


Gold nanoparticles were generated over SPCEs according to previous publications [23,24]. Briefly, 40 µL of a 1 mM AuCl4

- solution in 0.1 M HCl were dropped over the electrode surface and subjected to a constant current of -100 mA for 180 s. After gold deposition, the electrode surface was generously rinsed with ultrapure water and dried at room temperature before use. The characterization of SPCnAuEs and the study of the electrochemical behavior of mercury were carried out in a previous publication [24] and no further discussion will be included in the present work. For the voltammetric analysis of mercury, 5 µL of the upper acidic aqueous phase, obtained after microvolume back-extraction, were mixed with 37 µL of 0.5 M NaOH to obtain a suitable electrolytic medium [23]. A volume of 40 µL of this solution was dropped onto SPCnAuEs for SWASV experiments. Under optimized conditions [23,24], mercury was preconcentrated by applying a constant potential of +0.3 V for 240 s. Thereafter, the potential was recorded between +0.3 V and +0.55 V at a frequency of 80 Hz, amplitude of 30 mV and step potential of 4 mV. An anodic peak corresponding to the reoxidation of mercury appears at approximately +0.42 V and the height of this peak was employed to quantify the analyte. All experiments were carried out at room temperature and SPCnAuEs were discarded after a single use.


3.1. Optimization of sample preparation

Variables affecting the proposed methodology were the amount of chelating agent, sample pH, volume of 4 M HCl acceptor solution during back-extraction, ionic strength, sample volume, IL amount, vortex-assisted IL-

Capítulo 4

258

DLLME extraction time and back-extraction time. Among them, the amount of chelating agent, sample pH and volume of 4 M HCl acceptor solution were thoroughly evaluated in a previous publication [23], and the results on their behavior are applicable to the present work. The optimum values obtained for the amount of chelating agent and volume of 4 M HCl solution (i.e., 40 µL of 2 mg mL-1 solution and 10 µL, respectively) were considered acceptable. Before each analysis, urine sample pH was measured and ranged between 5.5 and 6.7. Thus, pH adjustment was considered unnecessary in accordance with the previous publication [23].

The other five variables affecting the proposed method (i.e., ionic strength, sample volume, IL amount, vortex-assisted IL-DLLME extraction time and back-extraction time) were investigated and optimized with a one-at-a-time strategy. The experiments were carried out in triplicate, employing 10 µg L-1 of Hg2+ aqueous standard solutions. The height of the anodic peak corresponding to the reoxidation of mercury was used as analyte signal to evaluate the overall extraction efficiency under different conditions.

3.1.1. Effect of ionic strength

Considering the high salt content in urine samples compared to water samples [23,25], the effect of ionic strength was evaluated through the addition of different amounts of NaCl to Hg2+ aqueous standard solutions (i.e., 0, 10 and 35% w/v of NaCl). The results (not included) showed no differences between NaCl free and salty solutions, and therefore, ionic strength adjustment of urine samples was unnecessary.

3.1.2. Effect of sample volume

The effect of sample volume was examined by testing 4, 6, 8 and 10 mL of the Hg2+ solution. The results are shown in Figure 1. As expected, higher sample volume led to a higher response since greater sample volume implies a greater amount of analyte. Volumes above 10 mL were not considered due to the limited availability of biological samples such as urine. A volume of 10 mL was finally employed in subsequent experiments.


259

Figure 1. Effect of sample volume. Extraction conditions: Amount of chelating agent, 40 µL of 2 mg mL-1 solution; IL amount, 0.06 g; vortex-assisted IL-DLLME extraction time, 3 min; volume of 4 M HCl acceptor solution during back-extraction, 10 µL; back-extraction time, 9 min.

3.1.3. Effect of IL amount

In order to study the effect of IL amount on the performance of the proposed sample preparation protocol, different quantities of [Hmim][NTf2] were tested (i.e., 0.02, 0.04, 0.06, 0.08 and 0.10 g). As shown in Figure 2, and considering the error bars, non-significant effects were observed when the amount of IL was increased between 0.02 and 0.10 g. However, considering the importance of waste reduction and the ease of IL droplet manipulation, a compromise value of 0.06 g of IL was finally chosen for following experiments.

0

10

20

30

40

50

60

4 6 8 10

Peak

hei

ght (

µA)

Sample volume (mL)

Capítulo 4

260

Figure 2. Effect of IL amount. Extraction conditions: Sample volume, 10 mL; amount of chelating agent, 40 µL of 2 mg mL-1 solution; vortex-assisted IL-DLLME extraction time, 3 min; volume of 4 M HCl acceptor solution during back-extraction, 10 µL; back-extraction time, 9 min.

3.1.4. Effect of vortex-assisted IL-DLLME extraction time

Extraction time is an important variable in microextraction techniques. This is because extraction involves transferring the analyte from the sample solution to the extractant phase, which is time-dependent. Extraction time in vortex-assisted IL-DLLME is generally considered as the vortex agitation time, therefore, this variable was evaluated to obtain the best extraction performance. Times between 0.5 and 5 min were examined. As shown in Figure 3, analyte signal increased as extraction time increased during the first 3 min, whereas longer extraction times did not significantly affect extraction efficiency. Consequently, 3 min of vortex agitation was selected as optimum value for further experiments.

0

10

20

30

40

50

60

0.02 0.04 0.06 0.08 0.1

Peak

hei

ght (

µA)

Amount of [Hmim][NTf2] (g)


261

Figure 3. Effect of vortex-assisted IL-DLLME extraction time. Extraction conditions: Sample volume, 10 mL; amount of chelating agent, 40 µL of 2 mg mL-1 solution; IL amount, 0.06 g; volume of 4 M HCl acceptor solution during back-extraction, 10 µL; back-extraction time, 9 min.

3.1.5. Effect of back-extraction time

The influence of microvolume back-extraction time was also evaluated considering vortex agitation times between 2 and 10 min. According to Figure 4, extraction efficiency increased with time until 9 min. Then, the signal remained constant. Therefore, the optimum back-extraction time was considered to be 9 min.

0

10

20

30

40

50

60

0.5 1 2 3 4 5

Peak

hei

ght (

µA)

Vortex-assited IL-DLLME extraction time (min)

Capítulo 4

262

Figure 4. Effect of back-extraction time. Extraction conditions: Sample volume, 10 mL; amount of chelating agent, 40 µL of 2 mg mL-1 solution; IL amount, 0.06 g; vortex-assisted IL-DLLME extraction time, 3 min; volume of 4 M HCl acceptor solution during back-extraction, 10 µL.

3.2. Stability of urine samples

Biological samples are known to degrade over time after collection. Therefore, it is important to establish proper preservation treatments or storage conditions to ensure the validity of analysis. A fresh urine sample obtained from a healthy human volunteer was spiked with 10 µg L-1 of Hg2+ and divided into two portions. One portion was stored in the refrigerator at 4 ºC whereas the other was kept at room temperature (i.e., 21 ºC). In order to determine their stability, Hg2+ was extracted from each urine portion on successive days using the proposed method. As apparent in Figure 5, urine samples need to be stored at low temperature if the analysis is not carried out on the day of collection. In addition, samples stored at 4 ºC need to be analyzed within the first 2 days as Hg2+ determination is greatly affected by longer storage times

0

10

20

30

40

50

60

2 4 6 8 9 10

Peak

hei

ght (

µA)

Back-extraction time (min)


263

Figure 5. Influence of the urine storage temperature on Hg2+ determination.


Different matrix effects were found when analyzing different urine samples (Figure S1), therefore, standard addition calibration was employed to evaluate quality analytical parameters under optimized conditions. To this end, three calibration curves were constructed using standards of five concentration levels from 0 to 20 µg L-1. Good linearity was obtained with correlation coefficient values (r) ranging from 0.990 to 0.999. The Student's t-test was applied to assess the linearity [26] showing values ranging from 12.2 (r=0.990; N=5) to 35.8 (r=0.999; N=5), thus rejecting the null hypothesis of non-linear correlation for a 5% significance level and 3 degrees of freedom (t0.05,3=3.18). The sensitivity of the instrumental measurements estimated by the slope of standard addition calibration curves ranged from (0.89±0.08) to (1.26±0.07) µA µg-1 L. The enrichment factor (EF) of the proposed procedure was studied in three different urine samples. Since complexity matrix of urine hinders Hg2+ direct determination by SPCnAuEs, the EF was evaluated by the following strategy. On one hand, the proposed procedure was applied to three non-doped urine samples and the final aqueous extracts were spiked at 15 µg L-1 of Hg2+

0

5

10

15

20

25

30

35

40

0 1 2 3 4 5 6 7 8 9 10

Peak

hei

ght (

µA)

Time (days)

Room temperature (21 ºC)Refrigerator (4 ºC)

Capítulo 4

264

just before the electrochemical measurements. On the other hand, the proposed procedure was applied to the urine samples spiked at 15 µg L-1 level. Then, the signals obtained in both procedures were compared obtaining EFs between 20 and 31.

Figure S1. Calibration curves obtained with aqueous standards and urine samples at different spiking levels: 5 µg L-1 (urine sample 1); 10 µg L-1 (urine sample 2); and 15 µg L-1 (urine sample 3).

The limit of detection (LOD) was determined empirically, measuring progressively more diluted concentrations of the analyte. Thereby, three different urine samples were spiked with progressively lower Hg2+ amounts and analyzed using the proposed method. The LOD was the lowest concentration whose signal could be clearly distinguished from blank. Figure S2 shows the voltammograms employed to establish the LOD in different urine samples. The LOD values ranged between 0.5 and 1.5 µg L-1, which is significantly lower than the threshold level established by the World Health Organization (WHO) for the normal mercury content in urine (i.e., 10-20 µg L-

1). Additionally, LOD and limit of quantification (LOQ) were evaluated using the blank signals and their standard deviations. Values between 1.1 and 1.3 µg

0

10

20

30

40

50

60

70

80

90

0 5 10 15 20 25

Peak

hei

ght (

µA)

Concentration (µg L-1)

Aqueous sample

Urine sample 1

Urine sample 2

Urine sample 3


265

L-1 were obtained for the LOD, which confirm the results obtained by the empirical approach. LOQ ranged between 1.9 and 2.4 µg L-1.

Figure S2. Square-wave voltammograms used to determine LOD emprirically in different urine samples. LOD was, respectively: (a) 0.5 µg L-1; (b) 1 µg L-1; and (c) 1.5 µg L-1.

0.32 0.36 0.40 0.44 0.48 0.520

2

4

6

8

10

12C

urre

nt (

A)

Potential (V)

0.5 g L-1

0.2 g L-1

blank

1 g L-1

0.75 g L-1

blank

0.35 0.40 0.45 0.500

2

4

6

8

10

12

Cur

rent

(A

)

Potential (V)

(a)

(b)

Capítulo 4

266

Figure S2 (cont.). Square-wave voltammograms used to determine LOD emprirically in different urine samples. LOD was, respectively: (a) 0.5 µg L-1; (b) 1 µg L-1; and (c) 1.5 µg L-1.

The accuracy of the method was evaluated by analyzing a certified urine sample (REC-8848/Level II) containing 17.3±4.3 µg L-1 of Hg2+. The higher concentration of metals in this sample compared with common concentrations found in urine samples, resulted in a non-detectable amount of analyte in the final acidic aqueous extract. Consequently, the amount of APDC was increased to 0.02 g to provide the excess of chelating agent necessary to efficiently complex and extract metals, including Hg2+. In addition, the presence of higher amount of APDC complexes, with Hg2+ and other co-existing metals, implies the necessity of using stronger acid conditions for a complete metal release during back-extraction. Therefore larger volume and higher concentration of the HCl acceptor aqueous solution (i.e., 20 µL and 6 M, respectively) were required. Employing the above-described modifications, the certified urine sample was analyzed using the standard addition method, obtaining a recovery (i.e., trueness) of 87%, and a precision expressed as the standard deviation estimated using regression line [26] of 3 µg L-1. These results confirm that the method is able to determine Hg at trace levels in urine samples. It should be emphasized that, despite the slight modification to the method when analyzing

0.35 0.40 0.45 0.500

2

4

6

8

10

Cur

rent

(

Potential (V)

1.5 g L-1

1 g L-1

blank

(c)


267

the reference material, the concentration of metals in this sample was significantly higher than the usual concentration found in human urine samples, for which the applicability of the originally proposed method has been demonstrated (see Section 3.4.).

3.4. Analysis of urine samples

The proposed method was applied to determine Hg2+ in real urine samples taken from healthy volunteers. Samples were initially analyzed and their Hg2+ content was found to be below the LOD of the present method. Consequently, three different urine samples were spiked adding different known amounts of Hg2+, which ranged from 5 to 15 µg L-1, and analyzed using standard addition calibration. To prepare the standard addition calibration, five aliquots of 9800 µL of each spiked urine sample were placed in test tubes and subsequently spiked with 0, 50, 100, 150 and 200 µL of a 1 mg L-1 Hg2+ aqueous solution, to which 200, 150, 100, 50 and 0 µL of ultrapure water were added, respectively, in order to reach an equal final volume (10 mL). After that, the microextraction procedure was applied and final aqueous extracts were electrochemically analyzed. Figure 6 shows the voltammograms obtained from urine sample spiked at 5 µg L-1.

Capítulo 4

268

Figure 6. Square-wave voltammograms obtained during the analysis of a spiked urine sample at 5 µg L-1 by the standard addition method. Legend shows calibration standards concentration.

Analyte concentration was calculated by extrapolation, giving the results in Table 1. According to these results, there were non-significant differences between the concentrations added and those found in the three urine samples, with relative recoveries ranging between 97 and 100%. Therefore, non-significant matrix effects were found with the proposed methodology.

Table 1. Determination of Hg2+ in spiked human urine samples.

Added concentration (µg L-1)

Found concentration ± SD* (µg L-1)

Relative recoveries (CV)

5 5.0 ± 0.9 100 (18) 10 9.7 ± 0.9 97 (9) 15 15 ± 2 100 (13)

*SD: standard deviation of x-value estimated using regression line [26].

0.32 0.36 0.40 0.44 0.48 0.520

5

10

15

20

25

30

35

40

45

Cur

rent

(A

)

Potential (V)

20 g L-1

15 g L-1

10 g L-1

5 g L-1

0 g L-1


269

3.5. Comparison with other methods

Characteristics of previously reported electrochemical methods for Hg2+ determination in urine samples are summarized in Table 2 for comparative purposes. The LOD, obtained using both methodologies (i.e., empirical and statistically), is equal or even better than those obtained in previous publications [9-13] and it is low enough to satisfy the threshold level established by the WHO for normal mercury concentrations in urine. In addition, the present approach affords special advantages. Firstly, the proposed sample preparation method avoids acid digestion, thus representing a safer and more environmentally-friendly procedure, which avoids the use of concentrated acids. By avoiding acid digestion, the proposed sample preparation protocol does not require time-consuming cooling steps and reduces the risk of mercury losses. Secondly, to our knowledge, this is the first time that SPELs have been employed to determine mercury in urine samples. SPELs are inexpensive devices which can be disposed after a single use. Thus, memory effects and tedious cleaning steps of the electrode surface between measurements are avoided. In addition, SPELs are easy to handle and their modification with gold nanoparticles is simple and, moreover, stripping experiments are rapid.

Capítulo 4

270

Table 2. Comparison of different electrochemical methods for Hg2+ determination in urine samples.

Electrode Sample pretreatment Electroanalytical technique (deposition time) LOD Ref.

MBX-EIGE Acid digestion FI-DPASV

(10 min) 0.38 µg L-1* [9]

Au-RDE Acid digestion SWASV (2 min) 0.8 µg L-1 [10]

DTDO modified GCE Acid digestion DPASV (25 min) 6 µg L-1* [11]

EDTA-CPME Acid digestion SWASV (10 min) 0.1 µg L-1* [12]

Gold film electrode US-assisted acid digestion ASCP (5 min) 0.3 µg L-1 [13]

SPCnAuEs Vortex-assisted IL-DLLME and back-extraction

SWASV (4 min) 0.5-1.5 µg L-1 This

work * LOD for aqueous standard solutions.

MBX, 2-mercaptobenzoxazole; EIGE, epoxy impregnated graphite electrode; FI, flow-injection; DPASV, differential pulse anodic stripping voltammetry; Au-RDE, gold rotating disk electrode; DTDO, 1,8-bis(dodecylthio)-3,6-dioxaoctane; GCE, glassy carbon electrode; EDTA, ethylenediaminetetraacetic acid; CPME, conducting polymer modified electrode; US, ultrasounds; ASCP, anodic stripping chronopotentiometry.


271

4. Conclusions

SPCnAuEs have been successfully combined with vortex-assisted IL-DLLME and microvolume back-extraction methodologies to determine mercury in non-digested urine samples. This combination synergistically exploits the advantages of environmentally friendly miniaturized sample preparation with the rapid, inexpensive, sensitive and selective determination of mercury with SPCnAuEs. The applicability of the proposed method has been successfully demonstrated in urine samples, reaching a LOD low enough to satisfy the threshold limit established by the WHO for normal mercury content in human urine. Thus, this method is proposed for rapid assessment of mercury contained in the urine of workers exposed to a contaminated space, among other possible applications.

Acknowledgments

The authors would like to thank the Spanish Ministry of Science and Innovation (projects n. CTQ2011-23968 and CTQ2011-24560), Generalitat Valenciana (Spain) (projects n. ACOMP/2013/072, GVA/2014/096 and PROMETEO/2013/038) and University of Alicante (Spain) (project n. GRE12-45) for the financial support. E. Fernández also thanks Ministry of Education (FPU13/03125) for her FPU grant.

References

[1] K.L. Nuttall, Interpreting mercury in blood and urine of individual patients, Ann. Clin. Lab. Sci. 34 (2004) 235-250.

[2] World Health Organization, May 2015, <http://www.who.int/foodsafety/publications/risk-mercuryexposure/en/>.

[3] C. Fernandez, A.C.L. Conceiçao, R. Rial-Otero, C. Vaz, J.L. Capelo, Sequential flow injection analysis system on-line coupled to high intensity focused ultrasound: green methodology for trace analysis applications as demonstrated for the determination of inorganic and total

Capítulo 4

272

mercury in waters and urine by cold vapor atomic absorption spectrometry, Anal. Chem. 78 (2006) 2494-2499.

[4] W.T. Corns, P.B. Stockwell, M. Jameel, Rapid method for the determination of total mercury in urine samples using cold vapour atomic fluorescence spectrometry, Analyst 119 (1994) 2481-2484.

[5] C. Burrini, A. Cagnini, Determination of mercury in urine by ET-AAS using complexation with dithizone and extraction with cyclohexane, Talanta 44 (1997) 1219-1223.

[6] M.N. Uddin, N.M. Shah, M.A. Hossain, M.M. Islam, Copper and mercury in food, biological and pharmaceutical samples: spectrophotometric estimation as Cu(DDTC)2, Am. J. Anal. Chem. 5 (2014) 838-850.

[7] A.N. Anthemidis, G.A. Zachariadis, C.E. Michos, J.A. Stratis, Time-based on-line preconcentration cold vapour generation procedure for ultra-trace mercury determination with inductively coupled plasma atomic emission spectrometry, Anal. Bioanal. Chem. 379 (2004) 764-769.

[8] B.M.W. Fong, T.S. Siu, J.S.K. Lee, S. Tam, Determination of mercury in whole blood and urine by inductively coupled plasma mass spectrometry, J. Anal. Toxicol. 31 (2007) 281-287.

[9] R. Ye, S.B. Khoo, Continuous flow and flow injection stripping voltammetric determination of silver(I), mercury(II), and bismuth(III) at a bulk modified graphite tube electrode, Electroanalysis 9 (1997) 481-489.

[10] Y. Bonfil, M. Brand, E. Kirowa-Eisner, Trace determination of mercury by anodic stripping voltammetry at the rotating gold electrode, Anal. Chim. Acta 424 (2000) 65-76.

[11] M.-S. Won, Y.-J. Bae, S.-S. Lee, Y.-B. Shim, Determination of Hg2+ ions using electrodes modified with dithia-podands having different end alkyl chain lengths, Electroanalysis 13 (2001) 1003-1007.

[12] M.A. Rahman, M.-S. Won, Y.-B. Shim, Characterization of an EDTA bonded conducting polymer modified electrode: its application for the simultaneous determination of heavy metal ions, Anal. Chem. 75 (2003) 1123-1129.


273

[13] R.A.A. Munoz, F.S. Felix, M.A. Augelli, T. Pavesi, L. Angnes, Fast ultrasound-assisted treatment of urine samples for chronopotentiometric stripping determination of mercury at gold film electrodes, Anal. Chim. Acta 571 (2006) 93-98.

[14] J.P. Metters, R.O. Kadara, C.E. Banks, New directions in screen printed electroanalytical sensors: an overview of recent developments, Analyst 136 (2011) 1067-1076.

[15] F.W. Campbell, R.G. Compton, The use of nanoparticles in electroanalysis: an updated review, Anal. Bioanal. Chem. 396 (2010) 241-259.

[16] M. Rezaee, Y. Assadi, M.-R. Milani Hosseini, E. Aghaee, F. Ahmadi, S. Berijani, Determination of organic compounds in water using dispersive liquid-liquid microextraction, J. Chromatogr. A 1116 (2006) 1-9.

[17] E. Yiantzi, E. Psillakis, K. Tyrovola, N. Kalogerakis, Vortex-assisted liquid-liquid microextraction of octylphenol, nonylphenol and bisphenol-A, Talanta 80 (2010) 2057-2062.

[18] J. Regueiro, M. Llompart, C. Garcia-Jares, J.C. Garcia-Monteagudo, R. Cela, Ultrasound-assisted emulsification-microextraction of emergent contaminants and pesticides in environmental waters, J. Chromatogr. A 1190 (2008) 27-38.

[19] Q. Zhou, H. Bai, G. Xie, J. Xiao, Temperature-controlled ionic liquid dispersive liquid phase micro-extraction, J. Chromatogr. A 1177 (2008) 43-49.

[20] C. Yao, J.L. Anderson, Dispersive liquid-liquid microextraction using an in situ metathesis reaction to form an ionic liquid extraction phase for the preconcentration of aromatic compounds from water, Anal. Bioanal. Chem. 395 (2009) 1491-1502.

[21] M.A. Farajzadeh, M.R.A. Mogaddam, Air-assisted liquid-liquid microextraction method as a novel microextraction technique; application in extraction and preconcentration of phthalate esters in aqueous sample followed by gas chromatography-flame ionization detection, Anal. Chim. Acta 728 (2012) 31-38.

Capítulo 4

274

[22] M.J. Trujillo-Rodríguez, P. Rocío-Bautista, V. Pino, A.M. Afonso, Ionic liquids in dispersive liquid-liquid microextraction, TrAC Trends Anal. Chem. 51 (2013) 87-106.

[23] E. Fernández, L. Vidal, D. Martín-Yerga, M. del C. Blanco, A. Canals, A. Costa-García, Screen-printed electrode based electrochemical detector coupled with ionic liquid dispersive liquid-liquid microextraction and microvolume back-extraction for determination of mercury in water samples, Talanta 135 (2015) 34-40.

[24] D. Martín-Yerga, M.B. González-García, A. Costa-García, Use of nanohybrid materials as electrochemical transducers for mercury sensors, Sens. Actuators B 165 (2012) 143-150.

[25] L. Vidal, A. Chisvert, A. Canals, A. Salvador, Sensitive determination of free benzophenone-3 in human urine samples based on an ionic liquid as extractant phase in single-drop microextraction prior to liquid chromatography analysis, J. Chromatogr. A 1174 (2007) 95-103.

[26] J.N. Miller, J.C. Miller, Statistics and Chemometrics for Analytical Chemistry, fifth ed., Pearson Prentice Hall, London, United Kingdom, 2005.


275

Rapid determination of hydrophilic phenols in olive oil by vortex-assisted reversed-phase dispersive liquid-liquid microextraction and screen-printed carbon electrodes


Departamento de Química Analítica, Nutrición y Bromatología e Instituto Universitario de Materiales, Universidad de Alicante, P.O. Box 99, E-03080 Alicante, Spain.

Abstract

A novel approach is presented to determine hydrophilic phenols in olive oil samples, employing vortex-assisted reversed-phase dispersive liquid-liquid microextraction (RP-DLLME) for sample preparation and screen-printed carbon electrodes for voltammetric analysis. The oxidation of oleuropein, hydroxytyrosol, caffeic acid, ferulic acid and tyrosol was investigated, being caffeic acid and tyrosol selected for quantification. A matrix-matching calibration using sunflower oil as analyte-free sample diluted with hexane was employed to compensate matrix effects. Samples were analyzed under optimized RP-DLLME conditions, i.e., extractant phase, 1 M HCl; extractant volume, 100 µL; extraction time, 2 min; centrifugation time, 10 min; centrifugation speed, 4000 rpm. The working range showed a good linearity between 0.075 and 2.5 mg L-1 (r=0.998, N=7) for caffeic acid, and between 0.075 and 3 mg L-1 (r=0.999, N=8) for tyrosol. The methodological limit of detection was established at 0.022 mg L-1 for both analytes, which is significantly lower than average contents found in olive oil samples. The repeatability was evaluated at two different spiking levels (i.e., 0.5 mg L-1 and 2 mg L-1) and coefficients of variation ranged from 8 to 11% (n=5). The applicability of the proposed method was tested in olive oil samples of different quality (i.e., refined olive oil, virgin olive oil and extra virgin olive oil). Relative recoveries varied between 93% and 100% showing negligible matrix effects. Finally, fifteen samples were analyzed by the proposed method and a high correlation with the traditional Folin-Ciocalteu spectrophotometric

Capítulo 4

276

method was obtained. Thereafter, the concentrations of the fifteen oil samples were employed as input variables in linear discriminant analysis in order to distinguish between olive oils of different quality.

Keywords: reversed-phase dispersive liquid-liquid microextraction, screen-printed electrodes, hydrophilic phenols, olive oil samples.

1. Introduction

Virgin olive oil (VOO) has become an essential component of the Mediterranean diet, having unique nutritional and organoleptic properties. Unlike other refined vegetable oils, VOO is produced exclusively by mechanical and physical means (e.g., cold-pressing, filtration, decantation, centrifugation) thus avoiding the oxidative degradation of bioactive compounds [1].

The chemical composition of VOO can be classified in majority and minority components. Majority components include monounsaturated and polyunsaturated fatty acids, mainly oleic and linoleic acids [2]. Minority components comprise a wide variety of chemical compounds such as carotenoids, phenols, aliphatic and triterpenic alcohols, sterols and hydrocarbons [2,3]. Carotenoids and phenols are the main components responsible for the antioxidant activity exerted by VOO, although carotenoids are present in significantly lower amounts. Lipophilic phenols (e.g., tocopherols) can be found in other vegetables oils; however, hydrophilic phenols (also known as polyphenols) are typically found only in VOO [2,3]. Hydrophilic phenols play a key role in the oxidative stability and healthy properties of VOO (e.g., anti-inflammatory, chemopreventive, cardiovascular) [2,3]. In addition, these phenols contribute to sensory qualities, affecting the typically pungent and bitter tastes [2,3]. Many different compounds constitute the hydrophilic phenolic fraction, including phenolic acids, phenolic alcohols, hydroxyisochromans, flavonoids, lignans and secoiridoids. The qualitative and quantitative content of these compounds is strongly affected by different factors such as the geographical origin, environmental conditions, olive ripening, harvesting, extraction methods and storage conditions [4,5].


277

Many efforts have focused on the characterization and quantification of the hydrophilic phenolic fraction in VOO samples. Powerful techniques for the separation, identification and quantification of individual compounds include liquid chromatography (LC) coupled to ultraviolet, fluorescence, mass spectrometry (MS) or nuclear magnetic resonance detector, with LC-MS being the most frequently employed combination [3]. However, the traditional Folin-Ciocalteu method, based on the colorimetric determination of total polyphenols, is still very useful to estimate the antioxidant capacity of VOO with a simple procedure and low cost [6]. Also, alternative electrochemical methods have been developed with the same purpose [7-16].

Inherent properties of olive oil samples (e.g., hydrophobicity, viscosity, complex chemical composition) make sample treatments necessary before instrumental analysis. Traditionally, solid-phase extraction (SPE) and liquid-liquid extraction (LLE) have been employed to isolate hydrophilic phenols prior to LC [3], spectrophotometry [3], or electroanalysis [7-10]. Nevertheless, recently reversed-phase dispersive liquid-liquid microextraction (RP-DLLME) has been introduced as valuable and green alternative to replace the aforementioned tedious and time-consuming techniques [17]. RP-DLLME is based on the dispersion in tiny droplets of a few µL of an aqueous solution in the hydrophobic sample. The cloudy solution presents a great contact surface area between the donor and acceptor phases, thus enhancing extraction efficiency [17]. After extraction (lasting a few seconds or minutes), phases are separated by centrifugation and the enriched aqueous phase is retrieved for subsequent analysis. RP-DLLME has been employed prior to LC analysis to determine hydrophilic phenols in VOO previously [17-21]; however, to the best of our knowledge, this miniaturized extraction technique in conjunction with electrochemical analysis has not been proposed to date. Here we present for the first time an analytical method to assess the hydrophilic phenolic fraction in olive oils using RP-DLLME as sample preparation technique and screen-printed carbon electrodes (SPCEs) as electrochemical transducers. This association synergistically combines the advantages of RP-DLLME (i.e., speed and ease of use, low sample volume, reduced generation of wastes, ecological, high enrichment factors and affordability) with the rapid response, inexpensive instrumentation and portability of SPCEs. Electrochemical behavior of the

Capítulo 4

278

main hydrophilic phenols (i.e., oleuropein, hydroxytyrosol, caffeic acid, ferulic acid and tyrosol) was evaluated with SPCEs, and subsequently caffeic acid and tyrosol were selected as model compounds. Parameters affecting RP-DLLME were studied using a multivariate optimization strategy. The applicability of the proposed method was tested in olive oils of different quality. Finally, fifteen olive oil samples were analyzed using the proposed method and the results were compared with those obtained with the Folin-Ciocalteu spectrophotometric method. Thereafter, found concentrations by the proposed method were subjected to linear discriminant analysis (LDA) in order to distinguish between olive oils of different quality.


2.1. Reagents and samples

Oleuropein (≥98%), hydroxytyrosol (≥98%), caffeic acid (≥98%), ferulic acid (99%) and tyrosol (>99.5%) standards were obtained from Sigma-Aldrich (Steinheim, Germany). Stock solutions of individual compounds (1000 mg L-1) were prepared in LC grade acetone from Sigma-Aldrich and stored in amber glass vials in the freezer (i.e., -18 ºC). Working solutions were prepared daily by proper dilution of stock solutions in LC grade hexane from Sigma-Aldrich. Fuming HCl (37%) from Merck (Darmstadt, Germany) was employed to prepare HCl aqueous solutions. The ultrapure water (resistivity of 18.2 MΩ cm at 25 ºC) employed to prepare aqueous solutions was obtained with a Millipore Direct System Q5™ purification system from Ibérica S.A. (Madrid, Spain). Analytes were dissolved in aqueous 0.1 M HCl solutions to study their electrochemical behavior with SPCEs.

Folin-Ciocalteu reagent from Fluka (Steinheim, Germany) and Na2CO3 (99%) from Prolabo (Paris, France) were employed in Folin-Ciocalteu assays.

Sunflower oil and fifteen olive oil samples of different trademark and quality, namely “olive oil”, VOO and extra virgin olive oil (EVOO), were purchased in local supermarkets. It should be noticed that commercial oils labeled as “olive oil” consist of mixtures of refined olive oil (up to 90%) and VOO or EVOO. Hereafter, they will be named as refined olive oil (ROO) to


279

avoid confusion. Samples were stored in the dark at room temperature and opened just before use to prevent the oxidative degradation of target analytes.


A vortex mixer from Heidolph (Swabach, Germany) was used to assist RP-DLLME. A centrifuge from Selecta (Barcelona, Spain) was used for phase separation.

A Multi Autolab/M101 Potentiostat/Galvanostat from Metrohm Autolab B.V. (Utrecht, The Netherlands) controlled by NOVA software version 1.10 was used for electrochemical experiments. SPCEs (ref. DRP-110) with three-electrode configuration were purchased from DropSens (Oviedo, Spain). The working disk-shaped electrode, 4 mm in diameter, and the counter electrode were made of carbon ink whereas the pseudo-reference electrode was made of silver. Specific connectors obtained from DropSens (ref. DRP-DSC) were used to connect SPCEs to the potentiostat.

An ultraviolet-visible spectrophotometer from Thermo Scientific (Waltham, MA, USA) was employed in Folin-Ciocalteu assays.

2.3. RP-DLLME

Under optimized conditions, 5 mL of hexane standards or oil samples (1 or 0.150 g depending on the oil) diluted to 5 mL with hexane were placed in test tubes. Then, 100 µL of aqueous 1 M HCl solution were added and the mixture was shaken for 2 min using vortex agitation. Next, phases were separated by centrifugation for 10 min at 4000 rpm. The upper organic phase was carefully removed with a glass pipette and the remaining acidic aqueous phase (i.e., 40 µL) was retrieved with a syringe for final analysis by differential pulse voltammetry (DPV) using SPCEs. Figure 1 shows a scheme of the overall procedure.

Capítulo 4

280

Figure 1. Vortex-assisted RP-DLLME coupled with SPCEs.


Cyclic voltammetry was employed to investigate the electrochemical behavior of hydrophilic phenols with SPCEs. Potential was recorded between 0.0 V and +1.2 V at 100 mV s-1 scan rate.

DPV was employed as electroanalytical technique after RP-DLLME. An aqueous 0.1 M HCl standard solution containing 10 mg L-1 of caffeic acid and

Sample Aqueous extractant

phase addition

Centrifugation

Removing organic phase

Vortex-assisted RP-DLLME

Withdrawal the enriched aqueous phase

Voltammetric analysis with SPCEs

0

10

20

30

40

50

0.2 0.4 0.6 0.8 1

Cur

rent

(µA

)

Potential (V)


281

tyrosol was employed to optimize DPV parameters. Potential was recorded between +0.2 V and +1.1 V. Optimum DPV parameters were: 100 mV modulation amplitude; 10 mV step potential; 0.05 s modulation time and 0.5 s interval time.

SPCEs were always discarded after a single use. All experiments were carried out in triplicate and at room temperature (i.e., 21 ºC).

2.5. Folin-Ciocalteu method

Hydrophilic phenols were also determined spectrophotometrically by the Folin-Ciocalteu method for comparative purposes. The calibration curve was constructed using caffeic acid aqueous standards from 0 to 300 mg L-1 (N=5) in 1 M HCl. 40 µL of each standard solution was mixed with 200 μL of Folin-Ciocalteu reagent, 800 μL of 7.5% Na2CO3 aqueous solution and diluted up to 4 mL with deionized water. The mixture was manually shaken for a few seconds and, after a 2 h reaction in the dark at room temperature (i.e., 21 ºC), the absorbance was measured at 765 nm. Analytes were extracted from olive oil samples using RP-DLLME according to the procedure described in Section 2.3 and 40 µL of final acidic aqueous extracts were subjected to the colorimetric assay (i.e., mixed with 200 μL of Folin-Ciocalteu reagent and 800 μL of 7.5% Na2CO3 solution, diluted up to 4 mL and incubated for 2 h in the dark before spectrophotometric determination). The concentration of total hydrophilic phenols was finally expressed as mg of caffeic acid equivalents per Kg of oil (i.e., mgCAE Kg-1) considering the preconcentration factor of RP-DLLME procedure and sample dilution.


A multivariate optimization strategy was carried out to determine optimum conditions for RP-DLLME. The statistical software NEMRODW® ("New Efficient Methodology for Research using Optimal Design") from LPRAI (Marseille, France) was used to build the experimental design matrix and evaluate the results. The current peak of caffeic acid and tyrosol were individually used as response functions for optimization.

Capítulo 4

282

LDA was carried out using the Statgraphics statistical computer package “Statgraphics Plus 5.1.” (Warrenton, VA, USA). The concentration of caffeic acid equivalents and tyrosol found during the analysis of ROO, VOO and EVOO samples (expressed in mg Kg-1 of oil) were used as input variables during LDA.


3.1. Electroanalysis with SPCEs

3.1.1. Electrochemical behavior of hydrophilic phenols

Cyclic voltammograms of caffeic acid, hydroxytyrosol, oleuropein, ferulic acid and tyrosol are shown in Figure 2a. Caffeic acid, hydroxytyrosol and oleuropein (i.e., ortho-phenols) showed one anodic peak and one cathodic peak after reversing the scan direction. The reversibility of the oxidation reaction can be explained considering their chemical structure. These compounds possess two hydroxyl groups attached to a benzene ring in ortho position, which can be reversible oxidized to ortho-quinones. Ferulic acid showed one oxidation peak and one smaller and broader reduction peak on the negative-going scan. Although the mechanism underlying electrochemical oxidation of ferulic acid is still unclear, it is known to involve ortho-quinone moiety [22-24], whose reduction probably gave rise to the cathodic peak observed in the ferulic acid voltammogram. Finally, tyrosol showed a clearly irreversible process with one anodic peak, corresponding to the oxidation of the only hydroxyl group attached to the benzene ring, but no cathodic peak (Figure 2a). As also shown in Figure 2a, the oxidation of ortho-phenols occurred at very near potentials whereas mono-phenols were oxidized at higher and separated potentials.


283

3.1.2. Selection of model compounds

A 10 mg L-1 mixed standard solution containing all phenols under study was prepared in aqueous 0.1 M HCl and analyzed by DPV. Then, RP-DLLME was applied to a VOO sample under the following conditions: 100 µL of aqueous 0.1 M HCl as extractant phase, 3 min of extraction time and centrifugation for 10 min at 4000 rpm. After RP-DLLME, the final acidic aqueous extract was also analyzed by DPV. Figure 2b shows signals obtained with the mixed standard solution and the real sample after RP-DLLME for comparative purposes. As can be observed, ortho-phenols (i.e, oleuropein, hydroxytyrosol and caffeic acid) were simultaneously oxidized giving rise to an anodic peak at +0.5 V. At higher potential (i.e., +0.7 V), a peak was observed in the standard solution corresponding to ferulic acid oxidation, whereas this signal was almost negligible in the real sample. Finally, the oxidation peak of tyrosol was clearly distinguishable at +0.93 V in both voltammograms. Considering these results, caffeic acid was selected as reference compound to quantify total ortho-phenols as caffeic acid equivalents using the current peak at +0.5 V. Tyrosol was also included in subsequent experiments using the current peak at +0.93 V for quantification. On the contrary, ferulic acid was omitted in further investigations considering the low content of this compound in real samples.

Capítulo 4

284

Figure 2. (a) Cyclic voltammograms of 10 mg L-1 individual standards of hydrophilic phenols in aqueous 0.1 M HCl; and (b) DPV voltammograms of a 10 mg L-1 mixed standard solution in aqueous 0.1 M HCl (in red) and 0.1 M HCl aqueous extract after RP-DLLME of a VOO sample (in blue).

-20

-10

0

10

20

30

40

50

0 0.2 0.4 0.6 0.8 1 1.2

Cur

rent

(µA

)

Potential (V)

HydroxytyrosolCaffeic acidOleuropeinFerulic acidTyrosol

0

10

20

30

40

50

0.2 0.4 0.6 0.8 1

Cur

rent

(µA

)

Potential (V)

Ortho-phenols

Tyrosol

Ferulic acid

(a)

(b)


285

3.1.3. Study of interferences

The effect of interferences on the simultaneous electrochemical determination of caffeic acid and tyrosol was evaluated. To this end, 10 mg L-1 caffeic acid solutions in 0.1 M HCl containing different amounts of tyrosol (i.e., 0, 10, 30, 50 and 90 mg L-1) were analyzed by DPV. No effects were observed in the caffeic acid signal related to the presence of tyrosol (Figure S1). A previous publication reported an important effect of mono-phenols (i.e., phenol and tyrosol) on the electrochemical response of ortho-phenols (i.e., hydroxytyrosol) due to their adsorption on the electrode surface [10]. However, such an effect was not observed in our experiments with SPCEs.

Figure S1. DPV voltammograms of caffeic acid at different concentrations of tyrosol. The concentration of caffeic acid was maintained at 10 mg L-1 in all the experiments.

The effect of caffeic acid upon tyrosol signal was also investigated

analyzing 10 mg L-1 tyrosol solutions in 0.1 M HCl containing different amounts of caffeic acid (i.e., 0, 10, 30, 50 and 90 mg L-1). Tyrosol current peak was maintained constant in all tested solutions (Figure S2), revealing that

0

10

20

30

40

50

60

70

80

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1

Cur

rent

(µA

)

Potential (V)

Series2

Series3

Series4

Series5

Series1

50 mg L-1

90 mg L-1

30 mg L-1

10 mg L-1

0 mg L-1

Capítulo 4

286

neither caffeic acid nor its oxidation product (which is reversible reduced) blocked SPCEs surface.

Figure S2. DPV voltammograms of tyrosol at different concentrations of caffeic acid. The concentration of tyrosol was maintained at 10 mg L-1 in all the experiments.

Finally, we should mention that oxidation products of tyrosol were

adsorbed onto SPCEs surface as a second use of the same electrode after tyrosol determination provided a significantly lower electrochemical response. Thus, SPCEs were always discarded after a single use.

3.2. RP-DLLME multivariate optimization

Fractional factorial designs are employed for screening purposes when a large number of factors can affect extraction yield. One particular strategy is the Plackett-Burman design, which studies up to k=N-1 factors in N runs, where N is a multiple of 4 [25]. The Plackett-Burman design assumes that interaction between factors can be ignored so the main effects can be calculated

0

10

20

30

40

50

60

70

80

90

100

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1

Cur

rent

(µA

)

Potential (V)

Series2

Series3

Series5

Series10 mg L-1

90 mg L-1

50 mg L-1

30 mg L-1

10 mg L-1


287

with a reduced number of experiments, thereby saving time and resources. A Plackett-Burman design was used to construct the matrix of experiments, including five factors studied in eight runs. The five experimental factors selected at two levels were: HCl concentration, extractant volume, extraction time, centrifugation speed and centrifugation time. Table S1 shows the experimental factors and levels considered in the Plackett-Burman design. The eight experiments were randomly performed using 5 mL of hexane standards with 1 mg L-1 of caffeic acid and tyrosol. DPV was selected as electroanalytical technique. The peak heights of caffeic acid and tyrosol were separately employed as response functions.

Table S1. Experimental factors and levels of the Plackettt-Burman design.

The data obtained were analyzed by ANOVA and the results were visualized with the Pareto charts shown in Figure S3. The length of each bar was proportional to the influence of the corresponding factor, and the effects exceeding the reference vertical line can be considered significant with 95% of probability. In Figure S3a, the reference vertical line does not appear meaning that factors are far from the significance level. In addition, negative and positive signals reveal whether the system responses decrease or increase, respectively, when passing from the lowest to the highest level of the corresponding factor.

Factors Level

Low (-1) High (+1)

HCl concentration (M) 0.1 1

Extractant volume (µL) 100 200


Centrifugation speed (rpm) 2000 4000

Centrifugation time (min) 5 10

Capítulo 4

288

Figure S3. Pareto charts of the Plackett-Burman design obtained for: (a) caffeic acid; and (b) tyrosol.

As shown in Figure S3, none of studied factors had a significant effect on the system responses. However, extractant phase concentration and volume were the most important factors, having the same sign for both analytes studied here and, therefore, showing analogous behaviors during extraction. The positive effect of HCl concentration could be attributed to increased hydrogen-bonding interactions and thus, improved extraction performance. The negative effect of extractant volume can easily be explained considering that the smaller the volume of acceptor phase, the higher the concentration of the analyte in the extract. According to these results, HCl concentration was fixed at its highest level (i.e., 1 M) whereas extractant volume was fixed at the lowest level (i.e., 100 µL). The other factors were fixed at the most convenient experimental

(a)

(b)

HCl concentration (M)

Extractant volume (µL)


Centrifugation speed (rpm)

Centrifugation time (min)

HCl concentration (M)

Extractant volume (µL)


Centrifugation speed (rpm)

Centrifugation time (min)


289

level, namely: extraction time, 2 min; centrifugation speed, 4000 rpm; and centrifugation time, 10 min. Extraction time was fixed at its lowest level to reduce the length of the extraction procedure whereas centrifugation speed and time were fixed at the highest level to promote better phases separation during the analysis of olive oil samples. Further optimization was considered unnecessary since the limit of detection (LOD) of the proposed method was checked under the above mentioned conditions, being low enough to determine normal levels of hydrophilic phenols in olive oil.


Calibration curves were first constructed applying the proposed method to hexane standards of caffeic acid and tyrosol. However, important matrix effects were found when analyzing olive oil samples with relative recoveries ranging from 44% to 73%. Thus, matrix-matching calibration was proposed to correct matrix effects and evaluate quality analytical parameters. To this end, refined sunflower oil was employed as analyte-free sample matrix. Standards of 1 g of sunflower oil diluted up to 5 mL with hexane were subjected to the proposed method under optimized conditions. The concentration range studied was from 0.075 to 3 mg L-1 of oil and the final working range is shown in Table 1. Other main analytical parameters of the proposed method are also summarized in Table 1. The lowest concentration of working range was limited by the methodological limit of quantification (mLOQ), whereas the upper end for caffeic acid was established at 2.5 mg L-1 since the signals obtained with standards of 2.5 mg L-1 and 3 mg L-1 did not differ significantly. The resulting calibration curves possessed a high level of linearity (Table 1). The sensitivity was estimated by the slope of the calibration curves being (36.7±1.1) µA mg-1 L for caffeic acid and (26.4±0.8) µA mg-1 L for tyrosol. The repeatability of the proposed method, expressed as the coefficient of variation (CV), was evaluated by five consecutive extractions at concentrations of 0.5 and 2 mg L-1, ranging between 8 and 11% (Table 1). The enrichment factor (EF) of RP-DLLME was evaluated through the slope ratio of calibration curves with and without preconcentration (Table 1). Calibration curves without RP-DLLME were performed using caffeic acid and tyrosol standards in 1 M HCl (aqueous

Capítulo 4

290

acceptor phase solution), since the direct electrochemical determination of target analytes in sample solution was not feasible due to its organic nature.

LOD and LOQ were determined for the proposed method including RP-DLLME and electrochemical detection, therefore, they are referred to as methodological LOD (mLOD) and mLOQ, respectively [26]. mLOD was empirically determined measuring progressively more diluted concentrations of caffeic acid and tyrosol. mLOD was the lowest concentration whose signal could be clearly distinguished from blank, namely 0.022 mg L-1 for the two analytes under study. Additionally, the mLOD was statistically evaluated using three times the standard deviation of a sunflower oil standard solution containing very low concentrations of analytes. Obtained in this way, mLOD values were 0.006 mg L-1 and 0.003 mg L-1 for caffeic acid and tyrosol, respectively. The statistical estimation of the mLOD provided lower values than those obtained empirically. However, the empirical estimation is considered to provide much more realistic values and, therefore, the mLOD of the proposed method was established according to this approach. The mLOQ, defined as 3.3 times the mLOD [27], was 0.075 mg L-1. It should be noted that both mLOD and mLOQ were lower than the average content of hydrophilic phenols commonly found in olive oil samples [2].

In order to assess the accuracy (i.e., trueness and precision) of the method, three oil samples were subjected to recovery studies. Samples of ROO, VOO and EVOO were diluted up to 5 mL with hexane, with a dilution factor depending on the phenolic content. Thus, 1 g of sample was employed when analyzing ROO whereas lower amounts of VOO and EVOO (i.e., 0.150 g) were necessary to fit the range of concentrations studied in calibration curves. Diluted olive oil samples were analyzed by the proposed method using matrix-matching calibration. Thereafter, the diluted olive oil samples were spiked with caffeic acid and tyrosol at 0.5 mg L-1 and also analyzed by the proposed method using matrix-matching calibration. Table 2 summarizes the results obtained. Relative recoveries (i.e., trueness) ranged between 93 and 100%, whereas the precision of the method expressed as CV ranged between 4 and 10%. According to these results, we can conclude that matrix effects were not significant in the three selected oil samples using the proposed matrix-matching calibration strategy.


291

Table 1. Main analytical parameters of the proposed method obtained with the matrix-matching calibration using sunflower oil as analyte-free sample.

Analyte Working range (mg L-1) ra

CVb (%) mLODc (mg L-1)

mLOQd (mg L-1) EFe

0.5 mg L-1 2 mg L-1

Caffeic acid 0.075-2.5 0.998 (7) 11 10 0.022 0.075 38

Tyrosol 0.075-3.0 0.999 (8) 10 8 0.022 0.075 37

a Correlation coefficient: number of calibration points in parentheses. b Coefficient of variation: mean value for 5 replicated analysis of 0.5 and 2 mg L-1 spiked oil solutions. c Methodological limit of detection: experimentally obtained. d Methodological limit of quantification: calculated as 3.3 times the methodological limit of detection. e Enrichment factor: calculated as slope ratio between calibration curves with and without RP-DLLME.

Capítulo 4

292

Table 2. Concentrations added and found, relative recoveries and coefficients of variation (in parentheses) during recovery studies in different olive oil samples.

Caffeic acid Tyrosol

Added (mg L-1)

Found ± SDa (mg L-1)

Relative recoveryb (%)

Added (mg L-1)

Found ± SDa (mg L-1)

Relative recoveryb (%)

ROO 0 0.322 ± 0.015 - 0 0.343 ± 0.008 -

0.5 0.808 ± 0.018 97 (5) 0.5 0.84 ± 0.02 100 (4)

VOO 0 0.907 ± 0.002 - 0 0.928 ± 0.002 -

0.5 1.37 ± 0.03 93 (6) 0.5 1.42 ± 0.03 98 (6)

EVOO 0 1.629 ± 0.007 - 0 2.316 ± 0.007 -

0.5 2.12 ± 0.05 97 (10) 0.5 2.80 ± 0.05 97 (9)

aStandard deviation of three replicated analyses. bCoefficient of variation in parentheses.


293

3.4. Analysis of olive oil samples

Fifteen olive oil samples, including five ROOs, five VOOs and five EVOOs, were analyzed with the proposed method using using matrix-matching calibration (Section 3.3) As mentioned before, samples were diluted up to 5 mL with hexane, with a dilution factor depending on the phenolic content. Thus, 1 g of sample was employed when analyzing ROOs whereas lower amounts (i.e., 0.150 g) were employed when analyzing VOOs and EVOOs. Found concentrations were expressed as mg Kg-1 of oil considering the dilution factors and results are shown in Table 3. As expected, the lowest content of hydrophilic phenols corresponded to ROO samples whereas the highest concentrations were found in EVOO samples.

Table 3. Caffeic acid equivalents and tyrosol content found in fifteen olive oil samples of different quality analyzed by the proposed method.

Oil sample Caffeic acid equivalents (mg Kg-1)

Tyrosol (mg Kg-1)

ROO

1 4.2±0.7 5.2±0.05 2 4.32±0.04 4.60±0.10 3 5.1±0.5 4.58±0.15 4 10.5±0.8 3.87±0.05 5 6.3±0.5 5.38±0.08

VOO

1 63±6 44.8±1.6 2 64.4±1.9 54.0±1.8 3 51±8 55.1±1.5 4 60±5 47±4 5 27.2±0.8 41.2±1.6

EVOO

1 77±5 52.9±1.9 2 75±3 51±3 3 72±2 49.3±1.2 4 78±3 51.7±1.1 5 69±3 52±2

Capítulo 4

294

3.5. Comparison with other electrochemical methods

For comparative purposes the characteristics of previously reported electrochemical methods for hydrophilic phenols determination in olive oil samples are summarized in Table 4. As can be seen, most of the reported methods involve slow and tedious sample preparation procedures, consuming large amounts of reagents and organic solvents. In addition, some methods use home-made electrochemical devices and complex modifications of electrode surfaces, thus hindering their widespread laboratory use and reducing analysis throughput. By contrast, the proposed method combines a simple, fast and environmentally friendly sample preparation technique with electrochemical detection using unmodified, inexpensive and commercially available SPCEs, thus providing unique advantages. Finally, lower LOD values were obtained with the proposed method compared to those obtained in the previously reported works.


295

Table 4. Electrochemical methods for hydrophilic phenols determination in olive oil samples.

Electrode Sample preparation Electrochemical technique LOD Ref.

SPCE Extraction with glycine buffer 10 mM pH 2, NaCl

10 mM (oil:buffer, 1:10). Dilution of the final extract with glycine buffer (1:10)

DPV 0.25 mg Kg-1 [7]

Tyrosinase-based biosensor -

Amperometric monitoring of O2 consumption during phenols

oxidation reaction catalyzed by tyrosinase. FIA system

4 mg Kg-1 [7]

Array of CPE (five modified with

phthalocyanine derivatives, six modified with

polypirrole and one unmodified)

7 g of oil dissolved in hexane (10 mL) and extracted three times with 30 mL methanol:water

(60:40, v:v). Evaporation of the extract until dryness and reconstitution in 25 mL of 0.1 M KCl

aqueous solution

CV and SWV - [8]

SPGE Solid-phase extraction with C18 cartridge Amperometric detection in a FIA system - [9]

SPCE

25 g of oil dissolved in hexane (25 mL) and extracted three times with 15 mL methanol:water (3:2, v:v). Dilution of the final extract with ultra-

pure water up to 50 mL

SWV 1.25 mg Kg-1 [10]

FIA, flow injection analysis; CPE, carbon paste electrode; CV, cyclic voltammetry; SWV, square-wave voltammetry; SPGE; screen-printed graphite electrode.

Capítulo 4

296

Table 4 (cont.). Electrochemical methods for hydrophilic phenols determination in olive oil samples.

CPE, carbon paste electrode; CV, cyclic voltammetry; SWV, square-wave voltammetry; GCE, glassy carbon electrode; FIA, flow injection analysis; SDS, sodium dodecyl sulfate.

Electrode Sample preparation Electrochemical technique LOD Ref.

CPE modified with oils as electroactive

binder material - CV and SWV - [11,12]

GCE Oil dilution with chloroform containing 2% acetic

acid and 3.2% tetrabutylammonium bromide (oil: chloroform solution, 1:100)

Amperometric detection in a FIA system - [13]

GCE Preparation of oil-in-water nanoemulsions using Tween 20 and SDS in 100 mM acetate buffer

Amperometric detection in a FIA system 0.5 mg L-1 [14]

SPCE modified with polypyrrole

Emulsions preparation by sonicating 25 mL of 0.2 M SDS aqueous solution with 5 mL of oil sample for

15 min CV - [15]

Pencil-drawn paper-based carbon

electrode - CV - [16]

SPCE RP-DLLME DPV 0.022 mg L-1 This work


297

3.6. Comparison with the Folin-Ciocalteu method

The concentration of hydrophilic phenols found in the fifteen olive oil samples analyzed by the proposed method (Section 3.4) was expressed as mg of caffeic acid equivalents and tyrosol per Kg of oil sample (mgCAE+TY Kg-1). Then, samples were analyzed by the Folin-Ciocalteu method according to the procedure described in Section 2.5. A graphic comparison of the results of both procedures is shown in Figure 3.

Figure 3. Graphical comparison of the results obtained with the proposed method and the Folin-Ciocalteu method.

As can be observed, lower concentrations were systematically found with the proposed method compared to those obtained with the reference method. This outcome could be explained considering the following: firstly, the Folin-Ciocalteu method estimates the total polyphenol content whereas the proposed electrochemical method only reflects the concentration of ortho-phenols and tyrosol; secondly, the Folin-Ciocalteu reagent is considered a non-specific reagent by many authors since it can be reduced by non-phenolic compounds

y = (0.64±0.03)x + (7±4)

0

50

100

150

200

0 50 100 150 200 250

Pro

pose

d m

etho

d (m

g CA

E+TY

Kg-1

)

Folin-Ciocalteu (mgCAE Kg-1)

r=0.983

Capítulo 4

298

[28]. Thus, the Folin-Ciocalteu method could also reflect the presence of other oxidizable species present in the sample extract. Despite these differences, Figure 3 shows a high correlation between the results obtained by the two methods. Accordingly, to estimate the antioxidant capacity of olive oil samples, we can conclude that RP-DLLME coupled to electrochemical detection with SCPEs is a valuable alternative to Folin-Ciocalteu. Finally, we should point out that electrochemical determination with SPCEs enables us to distinguish ortho-phenols from tyrosol (i.e., mono-phenol), whereas a colorimetric method other than Folin-Ciocalteu is required to do so [3]. Therefore, the proposed method possesses unique advantages as it is simple, easy to handle and less-time consuming, given it does not require the incubation time (i.e., 2 h) inherent to colorimetric reactions.

3.7. Discriminant analysis

LDA was selected to assess the capability of the proposed method to distinguish olive oil samples of different quality. LDA is a supervised classification method whose main objective is to find a rule for allocating a new object of unknown group to the correct group, using a number of objects whose group membership is known [27]. With this aim, LDA maximizes the variation between pre-specified groups and minimizes the variation within a group, by the condensation of original variables into a set of orthogonal functions (i.e., linear discriminant functions, LDFs) with a minimum loss of information [27]. Thereby, the number of orthogonal LDFs is equal to the number of groups minus one.

LDA analysis was applied in order to find a predictive classification model able to separate olive oil samples according to their quality in three main groups, namely ROO, VOO and EVOO. Figure 4 shows the graphical representation of the LDFs of the obtained classification model. As can be seen, LDF-1 possessed a higher discriminant capacity than LDF-2 since it completely separated ROO from EVOO and VOO, whereas LDF-2 may help in separating the latter two types of olive oils. The higher discrimination capacity of LDF-1 was also revealed by the percentage of variance, being 97% for this


299

function. Nevertheless, both LDFs possessed a p-value lower than 0.05 revealing their statistical significance with 95% probability.

Figure 4. Graphical representation of LDFs obtained during LDA. Groups are shown with different symbols: ROO, squares; VOO, triangles; EVOO, circles. Crosses mark the centroid of each group.

The success of LDA at allocating oil samples correctly was tested using three samples of different quality. ROO and EVOO were correctly classified. However, VOO was classified as EVOO as a consequence of the overlap of this two observed in Figure 4. According to these results, we can conclude that the proposed procedure is able to distinguish between ROO and olive oils of higher quality (i.e., VOO, EVOO) and could be used to detect adulterations.

4. Conclusions

For the first time, RP-DLLME has been successfully combined with SPCEs to determine hydrophilic phenols in olive oil samples. Thereby, the advantages of miniaturized systems, both in sample preparation and detection stage, have been synergistically exploited. On the one hand, RP-DLLME involves a fast and easy-to-handle procedure with a significantly low consumption of organic solvents compared to SPE or LLE techniques, thus

OLIVA

VIRGEN

VIRGEN EXTRA

Centroides

Gráfica de Funciones Discriminantes

-9 -6 -3 0 3 6

Función 1

-3.2

-2.2

-1.2

-0.2

0.8

1.8

Funció

n 2

LDF-1

LDF-

2

-9 -6 -3 0 3 6

1.8

0.8

-0.2

-1.2

-2.2

-3.2

Capítulo 4

300

making it environmentally friendly. On the other hand, unmodified and commercially available SPCEs provide a rapid and sensitive response with affordable and portable instrumentation.

The multivariate optimization strategy used here enabled us to rapidly and economically establish RP-DLLME operation conditions. The matrix-matching calibration using refined sunflower oil as analyte-free sample resulted in a simple and suitable strategy to compensate matrix effects. The proposed method provided results that closely correlate with the well-established Folin-Ciocalteu method, which are useful to predict the results provided by time-consuming colorimetric assays. In addition, the proposed method is simpler, more time-efficient and enables us to distinguish ortho-phenols from tyrosol. Finally, the proposed method in combination with LDA has resulted in a suitable strategy to discriminate between ROO and higher quality olive oils. Therefore, RP-DLLME coupled to SPCEs is a novel and promising alternative to determine hydrophilic phenols in olive oil samples, is affordable for any laboratory and has a potential application for the rapid assessment of olive oil quality and detect fraudulent practices (e.g., adulterations).

Acknowledgements

The authors would like to thank Spanish Ministry of Science and Innovation (project no. CTQ2011-23968) and Generalitat Valenciana (Spain) (projects nos. GVA/2014/096 and PROMETEO/2013/038) for the financial support. E. Fernández also thanks Spanish Ministry of Education for her FPU grant (FPU13/03125).

References

[1] M.Z. Tsimidou, D. Boskou, The health claim on “olive oil polyphenols” and the need for meaningful terminology and effective analytical protocols, Eur. J. Lipid Sci. Technol. 117 (2015) 1091-1094.

[2] M. Servili, B. Sordini, S. Esposto, S. Urbani, G. Veneziani, I. Di Maio, R. Selvaggini, A. Taticchi, Biological activities of phenolic compounds of extra virgin olive oil, Antioxidants 3 (2014) 1-23.


301

[3] A. Segura-Carretero, A. Carrasco-Pancorbo, A. Bendini, L. Cerretani, A. Fernández-Gutiérrez, Analytical determination of polyphenols in olive oil, in: V.R. Preedy, R.R. Watson (Eds.), Olives and Olive Oil in Health and Dessease Prevention, 1st ed., Elsevier, London (UK), 2010: pp. 509-523.

[4] D. Krichene, M.D. Salvador, G. Fregapane, Stability of virgin olive oil phenolic compounds during long-term storage (18 months) at temperatures of 5-50°C, J. Agric. Food Chem. 63 (2015) 6779-6786.

[5] A.M. Gómez-Caravaca, R.M. Maggio, L. Cerretani, Chemometric applications to assess quality and critical parameters of virgin and extra-virgin olive oil. A review, Anal. Chim. Acta 913 (2016) 1-21.

[6] T. Gutfinger, Polyphenols in olive oils, J. Am. Oil Chem. Soc. 58 (1981) 966-968.

[7] C. Capannesi, I. Palchetti, M. Mascini, A. Parenti, Electrochemical sensor and biosensor for polyphenols detection in olive oils, Food Chem. 71 (2000) 553-562.

[8] M.L. Rodríguez-Méndez, C. Apetrei, J.A. de Saja, Evaluation of the 19 polyphenolic content of extra virgin olive oils using an array of voltammetric sensors, Electrochim. Acta 53 (2008) 5867-5872.

[9] M. Del Carlo, A. Amine, M. Haddam, F. della Pelle, G.C. Fusella, D. Compagnone, Selective voltammetric analysis of o-diphenols from olive oil using Na2MoO4 as electrochemical mediator, Electroanalysis 24 (2012) 889-896.

[10] T.A. Enache, A. Amine, C.M.A. Brett, A.M. Oliveira-Brett, Virgin olive oil ortho-phenols electroanalytical quantification, Talanta 105 (2013) 179-186.

[11] C. Apetrei, M.L. Rodríguez-Méndez, J.A. de Saja, Modified carbon paste electrodes for discrimination of vegetable oils, Sensors Actuators B 111-112 (2005) 403-409.

[12] C. Apetrei, F. Gutierez, M.L. Rodríguez-Méndez, J.A. de Saja, Novel method based on carbon paste electrodes for the evaluation of bitterness in extra virgin olive oils, Sensors Actuators B 121 (2007) 567-575.

[13] M.S. Cosio, D. Ballabio, S. Benedetti, C. Gigliotti, Evaluation of different storage conditions of extra virgin olive oils with an innovative

Capítulo 4

302

recognition tool built by means of electronic nose and electronic tongue, Food Chem. 101 (2007) 485-491.

[14] S. Benedetti, M.S. Cosio, M. Scampicchio, S. Mannino, Nanoemulsions for the determination of the antioxidant capacity of oils by an electrochemical method, Electroanalysis 24 (2012) 1356-1361.

[15] I.M. Apetrei, C. Apetrei, Voltammetric e-tongue for the quantification of total polyphenol content in olive oils, Food Res. Int. 54 (2013) 2075-2082.

[16] N. Dossi, R. Toniolo, F. Terzi, E. Piccin, G. Bontempelli, Simple pencil-drawn paper-based devices for one-spot electrochemical detection of electroactive species in oil samples, Electrophoresis 36 (2015) 1830-1836.

[17] P. Hashemi, F. Raeisi, A.R. Ghiasvand, A. Rahimi, Reversed-phase dispersive liquid-liquid microextraction with central composite design optimization for preconcentration and HPLC determination of oleuropein, Talanta 80 (2010) 1926-1931.

[18] P. Hashemi, F. Nazari Serenjeh, A.R. Ghiasvand, Reversed-phase dispersive liquid-liquid microextraction with multivariate optimization for sensitive HPLC determination of tyrosol and hydroxytyrosol in olive oil, Anal. Sci. 27 (2011) 943-947.

[19] M.P. Godoy-Caballero, M.I. Acedo-Valenzuela, T. Galeano-Díaz, New reversed phase dispersive liquid-liquid microextraction method for the determination of phenolic compounds in virgin olive oil by rapid resolution liquid chromathography with ultraviolet-visible and mass spectrometry detection, J. Chromatogr. A 1313 (2013) 291-301.

[20] P. Reboredo-Rodríguez, L. Rey-Salgueiro, J. Regueiro, C. González-Barreiro, B. Cancho-Grande, J. Simal-Gándara, Ultrasound-assisted emulsification-microextraction for the determination of phenolic compounds in olive oils, Food Chem. 150 (2014) 128-136.

[21] S. Dairi, T. Galeano-Díaz, M.I. Acedo-Valenzuela, M.P. Godoy-Caballero, F. Dahmoune, H. Remini, K. Madani, Monitoring oxidative stability and phenolic compounds composition of myrtle-enriched extra virgin olive during heating treatment by flame, oven and microwave using reversed phase dispersive liquid–liquid microextraction (RP-


303

DLLME)-HPLC-DAD-FLD method, Ind. Crops Prod. 65 (2015) 303-314.

[22] P. Hapiot, A. Neudeck, J. Pinson, H. Fulcrand, P. Neta, C. Rolando, Oxidation of caffeic acid and related hydroxycinnamic acids, J. Electroanal. Chem. 405 (1996) 169-176.

[23] R. Petrucci, P. Astolfi, L. Greci, O. Firuzi, L. Saso, G. Marrosu, A spectroelectrochemical and chemical study on oxidation of hydroxycinnamic acids in aprotic medium, Electrochim. Acta 52 (2007) 2461-2470.

[24] M. Salas-Reyes, J. Hernández, Z. Domínguez, F.J. González, P.D. Astudillo, R.E. Navarro, E. Martínez-Benavidez, C. Velázquez-Contreras, S. Cruz-Sánchez, Electrochemical oxidation of caffeic and ferulic acid derivatives in aprotic medium, J. Braz. Chem. Soc. 22 (2011) 693-701.

[25] D.C. Montgomery, Design and Analysis of Experiments, 7th ed., Wiley, New Jersey (USA), 2009.

[26] M.S. Epstein, Comparison of detection limits in atomic spectroscopic methods, in: L.A. Currie (Ed.), Detection in Analytical Chemistry: Importance, Theory, and Practice; American Chemical Society, Washington DC (USA), 1988; pp 109-125.

[27] J.N. Miller, J.C. Miller, Statistics and Chemometrics for Analytical Chemistry, 5th ed., Pearson Prentice Hall, London (UK), 2005.

[28] L.M. Magalhães, M.A. Segundo, S. Reis, J.L.F.C. Lima, Methodological aspects about in vitro evaluation of antioxidant properties, Anal. Chim. Acta 613 (2008) 1-19.

IV. Discusión general de los resultados/ General discussion of the results

Discusión general de los resultados

307

En esta memoria se ha tratado de desarrollar y evaluar nuevas estrategias para la miniaturización de la preparación de la muestra, así como avanzar en el conocimiento y la aplicación de otras ya existentes. Además, también se han propuesto y evaluado los SPELs como nuevos sistemas de detección electroquímica para la técnica de DLLME, debido a las importantes ventajas que aportan estos dispositivos frente a otras técnicas instrumentales tradicionalmente utilizadas.

En este apartado se pretende ofrecer un enfoque general de los resultados más relevantes obtenidos en esta memoria. Con este fin, por un lado se han tratado los aspectos relacionados con los elementos comunes a todos los capítulos, que son la miniaturización de la preparación de la muestra y el desarrollo de métodos analíticos sencillos, rápidos, económicos y ecológicos. Por otro lado, se han discutido los elementos diferenciadores entre los capítulos como son las técnicas de extracción y fases extractantes utilizadas, los analitos y muestras objeto de estudio, la optimización del proceso de extracción y, por último, los sistemas de separación y detección utilizados.

Elementos comunes de la memoria

308

IV.1. Elementos comunes de la memoria La miniaturización de la preparación de la muestra aporta numerosas

ventajas a la totalidad del proceso analítico. Por ello, uno de los objetivos principales de esta memoria ha sido desarrollar y evaluar nuevos elementos y estrategias que mejoren la etapa de preparación de la muestra siguiendo siempre la línea de la miniaturización.

En todos los trabajos de investigación presentados se ha utilizado una técnica miniaturizada de extracción. La Tabla IV.1 muestra, a modo de resumen, los datos más representativos de las diferentes estrategias de preparación de la muestra desarrolladas. De forma general, se puede afirmar que todas estas estrategias son:

Sencillas: debido a su simplicidad y fácil manejo. Rápidas: debido a los pocos minutos necesarios para su realización. Económicas: debido a la reducción del consumo de reactivos y

materiales. En todos los casos se han utilizado fases extractantes del orden de los miligramos o microlitros, y cantidades de muestra del orden de los gramos o mililitros.

Ecológicas: debido a la reducción del consumo de reactivos se ha conseguido disminuir la generación de residuos y desarrollar métodos analíticos menos perjudiciales para el medio ambiente y la salud del operador. Además, se han utilizado fases extractantes alternativas a los disolventes orgánicos tradicionales (i.e., ILs o disoluciones acuosas), modos de generar la dispersión en DLLME que suprimen la necesidad de utilizar disolventes orgánicos (i.e., energía de ultrasonidos, agitación con vórtex, reacción de metátesis), y se han incluido sistemas de detección de menor consumo energético que los sistemas tradicionales básados en las técnicas de GC o LC (i.e., SPELs).

Discusion general de los resultados

309

Tabla IV.1. Datos representativos de las diferentes estrategias de preparación de la muestra investigadas en esta memoria.

Capítulo Analitos Muestra Técnica de extracción Fase extractante Cantidad fase

extractante Elución o

re-extracción Cantidad muestra

Tiempo de extracción/elución

o re-extracción

1 BTEX Agua MSPE Zeolita/óxido de hierro 138 mg Elución con 0.5

mL de acetona 20 mL 11 min/5 min

2 Clorobencenos Agua HS-SDME [Emim]2[Co(NCS)4] 1 µL - 20 mL 10 min

3 Catecolaminas (DA, E, NE)

Orina EME 25 mM TFPBA en DEHPi

~20 µL Ácido fórmico 20 mM 0.5 mL 15 min

4

TNT Agua In-situ IL-DLLME [Hmim][NTf2] ~15 µL - 11 mL 0.5 min

Mercurio Agua In-situ IL-DLLME

[Hmim][NTf2] ~20 µL Re-extracción con 10 µL de

HCl 4 M 4 mL 0.5 min/14 min

Mercurio Orina VA-IL-DLLME

[Hmim][NTf2] 60 mg Re-extracción con 10 µL de

HCl 4 M 10 mL 3 min/9 min

Fenoles hidrofílicos

Aceite RP-

DLLME HCl 1 M 100 µL - 0.150-1 g 2 min


310

Dentro del Capítulo 4 se han incluido cuatro trabajos de investigación diferentes debido a las importantes características comunes que poseen. Este capítulo se ha centrado en el estudio de la técnica de DLLME, en diferentes modalidades, y su combinación con SPELs como sistemas de detección electroquímica. Este acoplamiento ha permitido combinar sinérgicamente las ventajas de utilizar sistemas miniaturizados tanto en la preparación de la muestra como en la etapa de detección. Por un lado, la DLLME minimiza el consumo de reactivos y de muestra, así como la generación de residuos. Además, esta técnica permite obtener altos factores de enriquecimiento en tiempos cortos de extracción y con un procedimiento bastante sencillo. Por otro lado, los SPELs ofrecen una respuesta rápida y sensible con una instrumentación de bajo coste y fácil manejo. Además, cabe mencionar que la detección electroquímica con SPELs también supone una reducción del consumo de disolventes con respecto a otras técnicas tradicionalmente acopladas a la DLLME, como por ejemplo las basadas en la LC. Como principal inconveniente de la determinación mediante SPELs podemos mencionar la necesidad, en ciertas ocasiones, de re-extraer el analito a un medio electrolítico adecuado para su determinación electroquímica. Sin embargo, esta re-extracción se realiza en unos pocos minutos y utilizando disoluciones acuosas de unos pocos microlitros (Tabla IV.1). Por tanto, a pesar de tratarse de una etapa adicional que idealmente debería eliminarse, dicha etapa no invalida las principales ventajas inherentes al acoplamiento DLLME-SPELs (e.g., rapidez, bajo coste, bajo consumo de reactivos, etc.). Los métodos propuestos basados en dicho acoplamiento suponen una alternativa muy interesante y prometedora frente a otros desarrollados para el mismo fin en los que se incluyen tratamientos de la muestra lentos y tediosos e instrumentación analítica de elevado precio y voluminosa. Además, aunque los métodos desarrollados incluyen ciertos elementos que limitan su portabilidad (e.g., atmósfera de nitrógeno, centrifugación), todos pueden ser considerados como un importante avance en el desarrollo de sistemas portátiles y económicos al alcance de cualquier laboratorio.


311

La Tabla IV.2 muestra un resumen de los principales parámetros analíticos obtenidos en los trabajos recogidos en esta memoria. Como puede observarse, en todos los casos se ha obtenido una buena linealidad en el intervalo de concentraciones señalado. Para evaluar este parámetro, se ha utilizado la inspección visual de las curvas de calibración derivadas de los resultados experimentales y sus coeficientes de correlación, que se encuentran comprendidos entre 0.950 y 0.9990. Además, en los Capítulos 1, 2 y 3, y en el Apartado III.4.3 se ha incluido un test estadístico (Student´s t-test) cuya utilización es muy recomendada para evaluar la linealidad ya que, además de considerar el coeficiente de correlación, tiene en cuenta el número de puntos de calibración. Por otro lado, en la Tabla IV.2 se observa como la precisión de los métodos desarrollados, expresada a través del CV, oscila entre el 3 y el 20%, obteniendo este valor máximo para el análisis por adición de estándar de una muestra de orina certificada. Finalmente, los LODs obtenidos son del orden de los ng L-1 o µg L-1, lo que ha permitido la determinación a niveles traza de los diferentes analitos objeto de estudio en muestras de diversa naturaleza (i.e., medioambiental, biológica y de alimentos).


312

Tabla IV.2. Principales parámetros analíticos de los métodos desarrollados en esta memoria.

Capítulo Analito Intervalo lineal Coeficiente de correlación (puntos de calibración)

CV1 (%) (nivel de concentración) LOD2 EF

1

Benceno 1-100 µg L-1 0.993 (N=6) 9 (40 µg L-1) 0.3 µg L-1 7.6 Tolueno 10-100 µg L-1 0.991 (N=6) 8 (40 µg L-1) 3 µg L-1 7.0

Etilbenceno 10-100 µg L-1 0.990 (N=6) 9 (40 µg L-1) 3 µg L-1 6.5 m,p-xileno 10-75 µg L-1 0.997 (N=5) 9 (40 µg L-1) 3 µg L-1 1.7

o-xileno 10-100 µg L-1 0.950 (N=6) 11 (40 µg L-1) 3 µg L-1 0.6

2

1,3-DCB 0.05-5 µg L-1 0.999 (N=6) 14 (0.7 µg L-1) 7 (3 µg L-1) 6.0 ng L-1 -



1,3,5-TCB 0.05-5 µg L-1 0.990 (N=6) 6 (0.7 µg L-1) 10 (3 µg L-1) 6.3 ng L-1 -



1,2,4,5-TeCB 0.05-5 µg L-1 0.995 (N=6) 4 (0.7 µg L-1) 6 (3 µg L-1) 7.6 ng L-1 -


PeCB 0.05-5 µg L-1 0.996 (N=6) 8 (0.7 µg L-1) 5 (3 µg L-1) 6.3 ng L-1 -

1Coeficiente de variación obtenido: a partir de cinco extracciones consecutivas a los niveles de concentración señalados (Capítulo 1); a partir de tres extracciones consecutivas a los niveles de concentración señalados (Capítulo 2). 2Límite de detección obtenido utilizando la señal del blanco y tres veces su desviación estándar (Capítulos 1 y 2).


313

Tabla IV.2 (cont.). Principales parámetros analíticos de los métodos desarrollados en esta memoria.

Capítulo Analito Intervalo lineal Coeficiente de correlación (puntos de calibración)

CV1 (%) (nivel de concentración) LOD2 EF

3 DA 25-125 µg L-1 0.996 (N=5) 13 (50 µg L-1) 1.8 µg L-1 6.0

E 25-125 µg L-1 0.990 (N=5) 13 (50 µg L-1) 5.0 µg L-1 2.2

4

III.4.1 TNT 10-80 µg L-1 0.9990 (N=4) 7 (20 µg L-1) 5 (50 µg L-1)

7 µg L-1 300

III.4.2 Mercurio 0.5-10 µg L-1 0.997 (N=6) 13 (3 µg L-1) 13 (10 µg L-1)

0.2 µg L-1 25

III.4.3 Mercurio 5-35 µg L-1 0.990-0.999 (N=5) 20 (17.3 µg L-1) 0.5-1.5 µg L-1 20-31

III.4.4 Ácido cafeico 0.075-2.5 mg L-1 0.998 (N=7) 11 (0.5 mg L-1)

10 (2 mg L-1) 0.022 mg L-1 38

Tirosol 0.075-3.0 mg L-1 0.999 (N=8) 10 (0.5 mg L-1) 8 (2 mg L-1)

0.022 mg L-1 37

1Coeficiente de variación obtenido: a partir de cinco extracciones consecutivas a los niveles de concentración señalados (Capítulo 3 y Apartados III.4.1, III.4.2 y III.4.4); a partir de sXE del análisis por adición de estándar de una muestra de orina certificada (Apartado III.4.3). 2Límite de detección obtenido: utilizando tres veces la relación señal/ruido (Capítulo 3); utilizando la señal del blanco y tres veces su desviación estándar (Apartado III.4.1); utilizando la señal del blanco y cinco veces su desviación estándar (Apartado III.4.2); experimentalmente (Apartados III.4.3 y III.4.4).


314

Las recuperaciones relativas obtenidas durante el análisis de las muestras reales se presentan en la Tabla IV.3. Para las muestras reales de agua (Capítulos 1 y 2, y Apartados III.4.1 y III.4.2), las recuperaciones relativas fueron calculadas mediante el cociente entre las señales obtenidas con la muestra real y con un patrón de agua ultrapura, siendo ambas disoluciones previamente dopadas con la misma cantidad de analito. De acuerdo con los resultados obtenidos, i.e., recuperaciones relativas comprendidas entre el 82 y el 114%, se puede afirmar que no existen efectos de matriz significativos en las muestras de agua analizadas. Por otro lado, el procedimiento que se acaba de describir dio lugar a importantes efectos de matriz durante el análisis de las muestras reales de orina y de aceite (Capítulo 3, y Apartados III.4.3 y III.4.4). Por tanto, para el análisis de estas muestras más complejas se adoptaron dos estrategias diferentes: la calibración por adición de estándar para el análisis de las muestras de orina, y la calibración por ajuste de matriz para el análisis de las muestras de aceite. Con dichas estrategias, se compensaron los efectos de matriz obteniendo unas recuperaciones relativas muy satisfactorias que oscilaron entre el 91 y el 117%.


315

Tabla IV.3. Recuperaciones relativas obtenidas durante el análisis de muestras reales con los métodos desarrollados en esta memoria.

Capítulo Analitos Muestra RR1 (%) (nivel de concentración)

1 BTEX

Agua de grifo 85-94 (40 µg L-1)

Agua de río 111-114 (40 µg L-1)

Agua residual 93-113 (40 µg L-1)

2 Clorobencenos

Agua de grifo 92-110 (0.7 µg L-1) 88-104 (3 µg L-1)

Agua de estanque

82-92 (0.7 µg L-1) 90-113 (3 µg L-1)

Agua residual 84-114 (0.7 µg L-1) 87-101 (3 µg L-1)

3 DA

Orina 91 (50 µg L-1)

E 117 (50 µg L-1)

4

III.4.1 TNT Agua de grifo 114 (40 µg L-1) Agua residual 109 (40 µg L-1)

III.4.2 Mercurio

Agua de grifo 106 (1 µg L-1) 108 (7 µg L-1)

Agua embotellada

98 (1 µg L-1) 103 (7 µg L-1)

Agua de río 97 (1 µg L-1) 98 (7 µg L-1)

Agua residual 97 (1 µg L-1) 95 (7 µg L-1)

III.4.3 Mercurio Orina 1 100 (5 µg L-1) Orina 2 97 (10 µg L-1) Orina 3 100 (15 µg L-1)

III.4.4

Ácido cafeico Tirosol Aceite de oliva 97 (0.5 mg L-1)

100 (0.5 mg L-1) Ácido cafeico

Tirosol Aceite de oliva

virgen 93 (0.5 mg L-1) 98 (0.5 mg L-1)

Ácido cafeico Tirosol

Aceite de oliva virgen extra

97 (0.5 mg L-1) 97 (0.5 mg L-1)

1Recuperación relativa calculada: con el cociente de las señales obtenidas con la muestra real y con un patrón de agua ultrapura dopados con la misma concentración de analito (Capítulos 1 y 2, y Apartados II.4.1 y III.4.2); como la relación entre la concentración encontrada mediante análisis por adición de estándar y la concentración añadida a la muestra de orina (Capítulo 3 y Apartado III.4.3); como la relación de señales obtenidas con la muestra real y con un patrón de aceite de girasol dopados con la misma concentración de analito (Apartado III.4.4).

Elementos no comunes de la memoria

316

IV.2. Elementos no comunes de la memoria Las principales diferencias introducidas en los distintos trabajos de

investigación presentados en esta memoria consisten en la técnica de extracción y fase extractante utilizadas, los analitos y muestras objeto de estudio, la optimización del proceso de extracción y, por último, los sistemas de separación y detección empleados. En el siguiente esquema (Figura IV.1) se muestran de modo resumido todos estos elementos diferenciadores.


317

Figura IV.1. Esquema de los elementos diferenciadores entre los distintos trabajos de investigación presentados en esta memoria.


318

A lo largo de los diferentes capítulos se han estudiado distintas técnicas

miniaturizadas de extracción. En este sentido, podemos establecer una primera y clara diferencia entre el Capítulo 1 y el resto de capítulos. El Capítulo 1 incluye una técnica miniaturizada de extracción en fase sólida: la MSPE. Está técnica aporta numerosas ventajas a la preparación de la muestra ya que aprovecha las propiedades magnéticas de la fase extractante para su manipulación y evita etapas lentas de centrifugado o filtración para la separación de las fases. Por otro lado, en el resto de capítulos se han utilizado distintas técnicas miniaturizadas de extracción en fase líquida. En el Capítulo 2, la técnica seleccionada ha sido la HS-SDME que permite la extracción de especies volátiles o semivolátiles desde matrices sucias evitando posibles interferencias de otras especies disueltas en la muestra y el desprendimiento de la gota suspendida por el contacto directo entre las fases. El Capítulo 3 incluye un estudio fundamental sobre la EME de analitos altamente polares. La EME de especies polares resulta un reto importante para los investigadores debido a la baja afinidad de esas especies por atravesar la SLM de carácter hidrofóbico. Los transportadores tradicionales utilizados para promover la EME de analitos polares llevan asociados ciertos inconvenientes relacionados con el aumento de la conductividad eléctrica de la SLM, una corriente alta del sistema y una elevada inestabilidad. Por lo tanto, las investigaciones dirigidas a extender el conocimiento sobre la EME de compuestos polares resulta de vital importancia para ampliar el campo de aplicación de esta técnica. Por último, en el Capítulo 4 se ha investigado la técnica de DLLME en diferentes modalidades: in-situ IL-DLLME, VA-IL-DLLME y RP-DLLME. En todas ellas se ha evitado la utilización de un disolvente orgánico para generar la dispersión de la fase extractante en la muestra, lo que se ha conseguido mediante una reacción de metátesis (in-situ-IL-DLLME) o mediante la agitación con un dispositivo vórtex (VA-IL-DLLME y RP-DLLME). Este hecho confiere un carácter más ecológico a los métodos desarrollados. Por otro lado, cabe destacar la rapidez de la DLLME ya que, tal y como se recoge en la Tabla IV.1, los tiempos de extracción para esta técnica en sus diferentes modalidades son significativamente menores (0.5-3 min) que los necesarios para otras técnicas (10-15 min).


319

Otro aspecto diferenciador entre los capítulos es la naturaleza de la fase

extractante. En el Capítulo 1 se ha utilizado como sorbente sólido una zeolita sintética, concretamente la ZSM-5, decorada con partículas de Fe2O3 que le confieren las propiedades magnéticas necesarias para la MSPE. Este nuevo sorbente posee propiedades muy ventajosas a la hora de utilizarse como fase extractante, entre las que cabe destacar su elevada porosidad y área superficial, su bajo coste y fácil disponibilidad, su síntesis sencilla y la posibilidad de reutilización. Dentro de la búsqueda de nuevos materiales para la mejora de la preparación de la muestra, los MILs están adquiriendo cada día una mayor importancia. Como se ha comentado en la introducción, la utilización de los MILs en técnicas de LPME se encuentra todavía en sus inicios siendo la búsqueda de MILs hidrofóbicos uno de los principales retos a afrontar para el tratamiento de muestras acuosas. Sin embargo, en el Capítulo 2 de esta memoria se ha presentado una estrategia alternativa que consiste en la utilización de un MIL hidrofílico en la técnica de HS-SDME, en la que no existe un contacto directo entre las fases y, por tanto, se evitan las limitaciones impuestas por su miscibilidad. Además, la propiedades magnéticas del MIL han permitido la colocación y adhesión de la gota a un pequeño imán de Nd durante la extracción, proporcionándole una mayor estabilidad que con la suspensión de la gota en la punta de la aguja de una jeringa. En el Capítulo 3 se ha propuesto una nueva SLM basada en el agente complejante TFPBA disuelto en DEHPi. Con este sistema se ha conseguido, por primera vez, la EME de especies básicas altamente polares, mediante la formación de complejos selectivos y reversibles en la interfase muestra/SLM. Por último, en el Capítulo 4 se ha utilizado el IL [Hmim][NTf2] y una disolución acuosa de HCl 1 M como fases extractantes para la DLLME. Ambas fases suponen una alternativa más ecológica a los disolventes orgánicos clorados tradicionalmente utilizados en esta técnica. El IL seleccionado ha sido ampliamente utilizado en técnicas de extracción al poseer una elevada inmiscibilidad en agua y una buena extractabilidad de compuestos orgánicos e iones metálicos. Además, la conductividad eléctrica y viscosidad del [Hmim][NTf2] también han sido fundamentales para su utilización en el Apartado III.4.1, en el que el IL ha servido como fase extractante y, posteriormente, como medio electrolítico para la determinación electroquímica del TNT con SPELs.


320

En cuanto a los analitos determinados, primeramente podemos distinguir entre los de naturaleza orgánica (i.e., BTEX, clorobencenos, catecolaminas, TNT y fenoles hidrofílicos) y los de naturaleza inorgánica (i.e., mercurio). Todos ellos han sido seleccionados teniendo en cuenta su relevancia medioambiental, biológica o nutricional. Los BTEX, los clorobencenos, el TNT y el mercurio presentan características contaminantes, tóxicas y, en algunos casos, carcinogénicas, por lo que el control de su presencia en muestras de agua potable, superficial o residual resulta de vital importancia. Además, los clorobencenos se seleccionaron como analitos modelo debido a su buena extractabilidad en ILs, mientas que el TNT se eligió considerando también sus características electroactivas. Por otro lado, el mercurio presenta un elevado factor de bioacumulación por lo que su determinación en muestras de orina es muy interesante para el diagnóstico rápido de una intoxicación por mercurio y para la adopción de medidas preventivas frente a su exposición. Por otra parte, las catecolaminas DA, E y NE son compuestos de elevado interés clínico ya que desempeñan un papel importante como neurotransmisores y como señales hormonales en el sistema nervioso central y en el sistema medular simpatoadrenal. Además, la determinación de estos analitos en orina es clave para la detección de tumores neuroendocrinos (e.g., feocrocitomas y neuroblastomas). Teniendo en cuenta su importancia biológica y su elevada polaridad, las catecolaminas DA, E y NE fueron seleccionadas como analitos modelo para el estudio fundamental sobre la técnica de EME llevado a cabo en el Capítulo 3. Por último, los fenoles hidrofílicos fueron el objetivo de una de las investigaciones presentadas ya que son los responsables de las valiosas propiedades organolépticas, antioxidantes y quimiopreventivas de los aceites de oliva virgen y virgen extra.

Las muestras estudiadas se dividen en tres grupos: agua, orina y aceite. Las muestras de agua de diferente origen (i.e., grifo, embotellada, residual, río y estanque) han sido las más investigadas debido a la gran importancia que tiene este recurso natural en nuestra vida diaria. Además, los problemas de escasez de agua son cada vez más frecuentes por lo que resulta imprescindible controlar la calidad y los niveles de contaminación de la misma. Por otro lado, la muestras de orina permiten realizar estudios en una matriz más compleja que la del agua. Asimismo, la recogida de esta muestra biológica es no invasiva


321

pudiéndose obtener mayores cantidades que, por ejemplo, la sangre o el suero, por lo que la posibilidad de diagnosticar enfermedades mediante su análisis es altamente ventajoso. Por último, el aceite de oliva se ha convertido en un componente esencial de la dieta mediterránea y la investigación acerca de su composición, la cual afecta directamente a su calidad, propiedades organolépticas y nutricionales, es un campo muy activo en la actualidad.

La metodología general de trabajo seguida en la mayoría de las investigaciones presentadas ha consistido en tres etapas fundamentales: (i) optimización de los diferentes factores que afectan a la preparación de la muestra; (ii) obtención y evaluación de los principales parámetros de calidad del método propuesto; y (iii) aplicación del método a muestras reales de interés. La optimización de la preparación de la muestra se ha llevado a cabo mediante el método convencional paso a paso (Capítulo 3 y Apartado III.4.3) o mediante la optimización multivariante (resto de capítulos o apartados). Esta segunda estrategia implica un ahorro de tiempo y recursos y es capaz de mostrar la existencia de interacciones entre factores, por lo que ha sido la más utilizada a lo largo de esta memoria. En todos los casos, se ha realizado una etapa previa de cribado debido al gran número de factores experimentales que afectan a la extracción. Posteriormente, una vez conocidos los factores con mayor influencia, se ha realizado un CCD que ha permitido obtener con un número reducido de experimentos las condiciones óptimas para la extracción.

Para finalizar, en cuanto a los sistemas de separación y detección utilizados podemos establecer una clara diferenciación entre los tres primeros capítulos, en los que se ha hecho uso de técnicas instrumentales tradicionales (Tabla IV.4), y el Capítulo 4, en el que se han propuesto los SPELs como sistemas de detección alternativos (Tabla IV.5). El sistema de separación-detección utilizado en los Capítulos 1 y 2 ha sido la GC-MS. En ambos casos se ha trabajado en modo SIM para aumentar la sensibilidad del detector y minimizar las señales procedentes de la matriz. No obstante, en el Capítulo 1 el eluato ha sido inyectado directamente en el sistema cromatográfico mientras que, en el Capítulo 2, se ha utilizado la desorción térmica como sistema de introducción de muestras. Este sistema ha permitido el paso de los analitos al interior de la columna cromatográfica impidiendo al mismo tiempo la entrada


322

del MIL debido a su baja volatilidad. La técnica de LC ha sido empleada en el Capítulo 3 para la separación de las catecolaminas utilizadas como analitos modelo antes de su detección por UV o MS/MS. El detector de UV se utilizó durante la optimización de las condiciones de extracción, al tratarse de un detector más sencillo y asequible. Sin embargo, la complejidad de las muestras de orina y la necesidad de alcanzar LODs más bajos, hizo necesario el uso de un sistema de UHPLC-MS/MS para la evaluación del método propuesto. Por último, los SPELs presentados en el Capítulo 4 suponen una alternativa interesante a las técnicas cromatográficas de los capítulos anteriores, las cuales son enormemente más lentas e incluyen una instrumentación voluminosa y de elevado coste. Los dispositivos utilizados están basados en electrodos de trabajo de diferente naturaleza en función del analito a determinar. En este sentido, los SPGEs y los SPCEs comerciales utilizados para la determinación de TNT y de fenoles hidrofílicos, respectivamente, no fueron sometidos a ningún tipo de tratamiento previo o modificación superficial antes de su uso. Con estos dispositivos se han obteniendo unos parámetros analíticos (e.g., sensibilidad, LOD) muy satisfactorios y a un coste muy bajo. Sin embargo, la determinación de mercurio en muestras de agua y de orina ha exigido la modificación de SPCEs con nanopartículas de oro. Los SPCnAuEs resultantes combinan las ventajas que ofrece el oro como material electródico (i.e., elevada afinidad por el mercurio, reducción a potenciales más positivos) con la elevada superficie del material nanoparticulado, que potencia de manera significativa la respuesta analítica frente a un electrodo de oro másico.


323

Tabla IV.4. Sistemas de separación-detección y tiempos de análisis de los métodos desarrollados en los tres primeros capítulos de la memoria.

Tabla IV.5. Electrodos, técnicas electroanalíticas y tiempos de análisis de los métodos desarrollados en el Capítulo 4.

Capítulo Analito Muestra Sistema de separación-detección

Tiempo de análisis

1 BTEX Agua GC-MS 28 min

2 Clorobencenos Agua TD-GC-MS 45 min

3 DA, E, NE Orina HPLC-UV 20 min

DA, E Orina UHPLC-MS/MS 5.5 min

Apartado Analito Muestra SPEL (técnica)

Tiempo de análisis

III. 4.1 TNT Agua SPGE (DPV)

20 min de purga con N2 y barrido de potencial en

pocos segundos

III. 4.2 Mercurio Agua SPCnAuE (SWASV)

4 min de deposición y barrido de potencial en

pocos segundos

III. 4.3 Mercurio Orina SPCnAuE (SWASV)

4 min de deposición y barrido de potencial en

pocos segundos

III. 4.4 Fenoles

hidrofílicos Aceite SPCE

(DPV) Pocos segundos

Common elements of the overview

324

The main objective of this overview has been to develop and evaluate new strategies for the miniaturization of sample preparation, as well as to improve other previously developed. In addition, SPELs have been presented as alternative detection systems for DLLME technique, due to the important advantages that these devices offer compared to other traditional techniques.

This section offers a general perspective of the most relevant results obtained in this overview. With this aim, the common elements to all of the chapters, such as the miniaturization of sample preparation and the development of simple, rapid, inexpensive and environmentally friendly analytical methods, are dealt with in detailed. On the other hand, the non-common elements, such as the employed extraction techniques and extractant phases, analytes and samples, optimization strategies, and separation and detection systems, are also discussed.

General discussion of the results

325

IV.3. Common elements of the overview The miniaturization of sample preparation provides numerous advantages

to the whole analytical process. For this reason, one of the main objectives of this overview has been to develop and evaluate new elements and strategies that improve sample preparation, always considering miniaturization concept.

In all the research works presented, a miniaturized extraction technique has been investigated. The most representative parameters of the sample preparation strategies developed are summarized in Table IV.6. In general, the developed strategies share the following common characteristics:

• Simplicity: due to the ease of handling. • Rapidity: due to the few minutes needed for operation. • Economical: due to the reduced consumption of reagents and

materials. In all cases, miligrams or microliters of extractant phase and grams or milliliters of sample have been used.

• Ecological: due to the reduced consumption of reagents and, therefore, the reduced generation of wastes, safer and more environmentally friendly methods have been developed. In addition, alternative extractant phases to those traditional organic solvents (i.e., ILs or aqueous solutions), dispersion modes in DLLME that avoid organic solvents (i.e., metathesis reaction, ultrasounds energy, vortex agitation), and detection systems with lower consumption of energy than tradicional systems based on GC or LC (i.e., SPELs) have been employed.


326

Table IV.6. Representative parameters of the sample preparation strategies developed in the present overview.

Chapter Analytes Sample Extraction technique Extractant phase Extractant

phase amount Elution or

back-extraction Sample amount

Extraction/elution or back-extracción

time

1 BTEX Water MSPE Zeolite/iron oxide 138 mg Elution with 0.5 mL of acetone

20 mL 11 min/5 min

2 Chlorobenzenes Water HS-SDME [Emim]2[Co(NCS)4] 1 µL - 20 mL 10 min

3 Catecholamines

(DA, E, NE) Urine EME 25 mM TFPBA

in DEHPi ~20 µL Formic acid

20 mM 0.5 mL 15 min

4

TNT Water In-situ IL-DLLME

[Hmim][NTf2] ~15 µL - 11 mL 0.5 min

Mercury Water In-situ IL-DLLME

[Hmim][NTf2] ~20 µL Back-extraction into 10 µL of 4 M HCl

4 mL 0.5 min/14 min

Mercury Urine VA-IL-DLLME

[Hmim][NTf2] 60 mg Back-extraction into 10 µL of 4 M HCl

10 mL 3 min/9 min

Hydrophilic phenols

Oil RP-DLLME

HCl 1 M 100 µL - 0.150-1 g 2 min


327

In Chapter 4, four different research works have been included considering their important common features. This Chapter has focused on the study of DLLME technique, in different modalities, and its combination with SPELs as electrochemical transducers. Thus, the advantages of using miniaturized systems both in sample preparation and detection step are synergistically combined. On the one hand, DLLME is a simple and rapid miniaturized extraction technique leading to low solvent and sample consumption, reduced generation of wastes and high enrichment factors. On the other hand, SPELs are miniaturized electrochemical devices that offer a rapid and sensitive response with inexpensive and easy to use instrumentation. In addition, it should be mentioned that the electrochemical detection with SPELs also implies a lower consumption of solvents than other techniques traditionally coupled to DLLME, as for example LC. The major drawback of SPELs could be the back-extraction needed, in some cases, to transfer target analyte into a suitable electrolytic for electrochemical detection. However, back-extraction is performed in few minutes and using low-volume aqueous solutions (Table IV.6). Therefore, this additional step does not devalue the remarkable advantages associated to the DLLME-SPEL coupling (e.g., speed, low cost, low reagent consumption, etc.). This novel combination can be presented as an interesting alternative to those traditional methods that include slow and tedious sample preparation procedures and expensive and bulky analytical instrumentation. Finally, it should be pointed out that, although the developed methods include some elements hindering the portability (e.g., nitrogen atmosphere, centrifugation), all of them can be considered a step forward in the development of portable and economical systems available to any laboratory.

Table IV.7 shows a summary of the main analytical parameters of the methods developed in this overview. In all cases, a good linearity was obtained in the range of concentrations indicated. In order to assess this parameter, the visual inspection of the calibration curves and the correlation coefficients, which ranged from 0.950 to 0.9990, have been considered. In addition, in Chapters 1, 2 and 3, and Section III.4.3, Student´s t-test has been included because its use is highly recommended to evaluate the linearity due to it takes into account both correlation coefficient and the number of calibration points. Table IV.7 also shows the CV values of the developed methods, which ranged


328

between 3 and 20%. The maximum value was obtained during the analysis of a certified urine sample by standard addition calibration. Finally, LOD values were in the order of ng L-1 or μg L-1, thus, the determination of target analytes at trace levels in different types of samples (i.e., environmental, biological and food) was achieved.


329

Table IV.7. Main analytical figures of merit of the methods developed in the present overview.

Chapter Analyte Lineal interval

Correlation coefficient (calibration points)

CV1 (%) (concentration level) LOD2 EF

1

Benzene 1-100 µg L-1 0.993 (N=6) 9 (40 µg L-1) 0.3 µg L-1 7.6 Toluene 10-100 µg L-1 0.991 (N=6) 8 (40 µg L-1) 3 µg L-1 7.0

Ethylbenzene 10-100 µg L-1 0.990 (N=6) 9 (40 µg L-1) 3 µg L-1 6.5 m,p-xylene 10-75 µg L-1 0.997 (N=5) 9 (40 µg L-1) 3 µg L-1 1.7

o-xylene 10-100 µg L-1 0.950 (N=6) 11 (40 µg L-1) 3 µg L-1 0.6

2









PeCB 0.05-5 µg L-1 0.996 (N=6) 8 (0.7 µg L-1) 5% (3 µg L-1) 6.3 ng L-1 -

1Coefficient of variation obtained: from five consecutive extractions at indicated levels (Chapter 1); from three consecutive extractions at indicated levels (Chapter 2). 2Limit of detection obtained using blank signal and three times its standard deviation (Chapters 1 and 2).


330

Table IV.7 (cont.). Main analytical figures of merit of the methods developed in the present overview.

Chapter Analyte Lineal interval Correlation coefficient (calibration points)

CV1 (%) (concentration level) LOD2 EF

3 DA 25-125 µg L-1 0.996 (N=5) 13 (50 µg L-1) 1.8 µg L-1 6.0

E 25-125 µg L-1 0.990 (N=5) 13 (50 µg L-1) 5.0 µg L-1 2.2

4

III.4.1 TNT 10-80 µg L-1 0.9990 (N=4) 7 (20 µg L-1) 5 (50 µg L-1) 7 µg L-1 300

III.4.2 Mercury 0.5-10 µg L-1 0.997 (N=6) 13 (3 µg L-1) 13 (10 µg L-1) 0.2 µg L-1 25

III.4.3 Mercury 5-35 µg L-1 0.990-0.999 (N=5) 20 (17.3 µg L-1) 0.5-1.5 µg L-1 20-31

III.4.4 Caffeic acid 0.075-2.5 mg L-1 0.998 (N=7) 11 (0.5 mg L-1)

10 (2 mg L-1) 0.022 mg L-1 38

Tyrosol 0.075-3.0 mg L-1 0.999 (N=8) 10 (0.5 mg L-1) 8 (2 mg L-1) 0.022 mg L-1 37

1Coefficient of variation obtained: from five consecutive extractions at indicated levels (Chapter 3 and Sections III.4.1, III.4.2 and III.4.4); obtained from sXE for the analysis of a certified urine sample by standard addition calibration (Section III.4.3). 2Limit of detection obtained: using three times the signal to noise ratio (Chapter 3); using blank signal and three times its standard deviation (Section III.4.1); using blank signal and five times its standard deviation (Section III.4.2); experimentally (Sections III.4.3 and III.4.4).


331

Relative recoveries obtained during the analysis of real samples are presented in Table IV.8. For water samples (Chapters 1 and 2, and Sections III.4.1 and III.4.2), relative recoveries were calculated using the signal ratio between the real sample and an aqueous standard solution spiked at the same concentration level. As Table IV.8 shows, relative recoveries ranged from 82 to 114%, and therefore, it can be concluded that matrix effects were not significant for the analysis of selected water samples. Following the above mentioned procedure, important matrix effects were found when analyzing real urine and oil samples (Chapter 3, and Sections III.4.3 and III.4.4). Thus, two different strategies were adopted for the analysis of these more complex matrices: standard addition calibration for urine samples, and matrix-matching calibration for oil samples. Following these calibration strategies, matrix effects were compensated and relative recoveries ranged from 91 to 117%.


332

Table IV.8. Relative recoveries obtained during the analysis of real samples with the methods developed in this overview.

Chapter Analyte Sample RR1 (%) (concentration level)

1 BTEX Tap water 85-94 (40 µg L-1)

River water 111-114 (40 µg L-1) Wastewater 93-113 (40 µg L-1)

2 Chlorobenzenes

Tap water 92-110 (0.7 µg L-1) 88-104 (3 µg L-1)

Pond water 82-92 (0.7 µg L-1) 90-113 (3 µg L-1)

Wastewater 84-114 (0.7 µg L-1) 87-101 (3 µg L-1)

3 DA Urine 91 (50 µg L-1) E 117 (50 µg L-1)

4

III.4.1 TNT Tap water 114 (40 µg L-1)

Wastewater 109 (40 µg L-1)

III.4.2 Mercury

Tap water 106 (1 µg L-1) 108 (7 µg L-1)

Bottled water 98 (1 µg L-1) 103 (7 µg L-1)

River water 97 (1 µg L-1) 98 (7 µg L-1)

Wastewater 97 (1 µg L-1) 95 (7 µg L-1)

III.4.3 Mercury Urine 1 100 (5 µg L-1) Urine 2 97 (10 µg L-1) Urine 3 100 (15 µg L-1)

III.4.4

Caffeic acid Tyrosol Olive oil 97 (0.5 mg L-1)

100 (0.5 mg L-1) Caffeic acid

Tyrosol Virgin olive oil 93 (0.5 mg L-1) 98 (0.5 mg L-1)

Caffeic acid Tyrosol

Extra virgin olive oil

97 (0.5 mg L-1) 97 (0.5 mg L-1)

1Relative recovery calculted: using the signal ratio between the real sample and an aqueous standard solution spiked at the same concentration level (Chapters 1 and 2, and Sections III.4.1 and III.4.2); using the ratio between found concentration by standard addition calibration and added concentration to urine sample (Chapter 3 and Section III.4.3); calculated using the signal ratio between the real sample and a standard solution of sunflower oil spiked at the same concentration level (Section III.4.4).


333

IV.4. Non-common elements of the overview The main differences between the research works presented in this

overview are related to the extraction technique, extractant phase, analytes and samples, optimization strategies, and separation and detection systems. Figure IV.2 shows a scheme summarizing all these non-common elements.

Non-common elements of the overview

334

Figure IV.2. Chart of the non-common elements between the different works included in this overview.


335

Different miniaturized extraction techniques have been studied throughout the chapters of this overview. In this regard, a first and clear difference can be established between Chapter 1 and the other chapters. On the one hand, Chapter 1 includes the study of a miniaturized solid-phase extraction

technique: MSPE. This technique offers numerous advantages to sample preparation since the magnetic properties of the extractant phase are exploited facilitating the manipulation and avoiding time-consuming centrifugation or filtration steps for phases separation. On the other hand, different miniaturized

liquid-phase extraction techniques have been investigated in Chapters 2, 3 and 4. In Chapter 2, HS-SDME has been employed for the extraction of chlorobenzenes from water samples. This technique is intended for the extraction of volatile or semi-volatile analytes from dirty matrices, preventing interferences of other species dissolved in sample solution and the dislodgment of suspended drop because the direct contact between phases is avoided. Chapter 3 includes a fundamental study on the EME of highly polar analytes. The EME of polar compounds is a challenging task due to the low affinity of those species to cross the hydrophobic SLM. In addition, the use of traditional carriers to promote EME of polar analytes is associated with serious disadvantages related to the increase of the electrical conductance of the SLM and to the high system-current during application of the electrical potential. Thus, investigations addressed to extend the currently limited knowledge about suitable carriers or organic solvents for the effective and stable EME of polar compounds are of great interest. Finally, different modalities of DLLME (i.e., in-situ IL-DLLME, VA-IL-DLLME and RP-DLLME) have been investigated in Chapter 4. In all of them, the use of organic solvents to disperse the extractant phase into the sample has been avoided. The dispersion has been achieved by a metathesis reaction (in-situ-IL-DLLME) or by vortex agitation (VA-IL-DLLME and RP-DLLME), providing a more ecological character to the developed methods. In addition, it is worth noting that extraction times in the different modalities of DLLME were significantly lower (0.5-3 min) than those employed with the other extraction techniques (10-15 min) (Table IV.6).

Another non-common element between the developed methods is the extractant phase. In Chapter 1, a synthetic zeolite, namely ZSM-5, decorated with Fe2O3 particles has been used as solid sorbent for MSPE. This new


336

sorbent possesses many remarkable properties such as high porosity and surface area, low cost, easy availability, simple synthesis and reutilization options. Regarding new materials to improve sample preparation, MILs are becoming increasingly important. As discussed in the introduction of this overview, the use of MILs in LPME techniques is in very preliminary stages and investigations about suitable hydrophobic MILs for the extraction from aqueous samples is one of the main challenges to face. However, an alternative strategy has been presented in Chapter 2, based on the use of a hydrophilic MIL for HS-SDME. With the proposed strategy, the direct contact between sample solution and extractant phase is avoided and, therefore, also the limitations imposed by their miscibility. In addition, the magnetic properties of the MIL have allowed disposing the drop into a small Nd magnet during extraction, providing a greater stability than with the suspension in the needle tip of a syringe. In Chapter 3, a new SLM based on the complexing agent TFPBA dissolved in DEHPi has been presented. With the proposed system, the EME of highly polar basic analytes has been achieved for the first time using the formation of selective and reversible complexes at the sample/SLM interface. In Chapter 4, the IL [Hmim][NTf2] and 1 M HCl aqueous solution have been employed as extractant phases for DLLME. Both phases represent a more ecological alternative to the traditionally used chlorinated organic solvents. The IL [Hmim][NTf2] has been widely used in extraction techniques due to its high immiscibility with water and good extractability of organic compounds and metals. In addition, the electrical conductivity and viscosity of [Hmim][NTf2] have been essential in Section III.4.1, in which the IL served both as extractant phase and electrolytic medium for the electrochemical determination of TNT with SPGEs.

Both organic (i.e., BTEX, chlorobenzenes, catecholamines, TNT and hydrophilic phenols) and inorganic (i.e., mercury) analytes have been determined in this overview. All of them have been selected considering their environmental, biological or nutritional relevance. BTEX, chlorobenzenes, TNT and mercury possess contaminant, toxic and, some of them, carcinogenic properties, therefore, controlling their presence in water samples is highly important. In addition, chlorobenzenes were selected as model analytes due to their good extractability in ILs, whereas TNT was chosen also considering the


337

electroactive characteristics. Mercury has a high bioaccumulation factor, so its determination in urine samples is very interesting for the rapid assessment of mercury intoxication and for the adoption of preventive measures for people exposed to a contaminated space. The catecholamines DA, E, and NE are compounds of clinical interest since they play a key role as neurotransmitters and hormones in the central nervous and sympathoadrenal systems. In addition, the determination of these analytes in urine is essential for the diagnosis of neuroendocrine tumors (e.g., pheochromocytoma and neuroblastoma). Considering their biological importance and high polarity, DA, E, and NE were selected as model analytes for the fundamental study about EME carried out in Chapter 3. Finally, hydrophilic phenols were the objective of one of the presented investigations since they are responsible of the valuable organoleptic, antioxidant and chemopreventive properties of virgin and extra virgin olive oils.

The samples investigated can be split into three main groups: water, urine and oil. Water samples from different sources (i.e. tap, bottled, residual, river and pond) have been the object of most studies due to the great importance that this natural resource has in our daily life. In addition, water is becoming an increasingly limited resource and it is important to control its quality and contamination levels. On the other hand, urine samples enable us to carry out studies in a more complex matrix than water. Additionally, the collection of this biological sample is non-invasive and greater amounts than, for example, blood or serum can be obtained. Thus, attempts to diagnose diseases through urine analysis are highly advantageous and interesting. Finally, olive oil has become an essential component of the Mediterranean diet and research focus on olive oil composition, which directly affects quality, organoleptic and nutritional properties, are a very active research field nowadays.

The general methodology followed in most of the presented works has been based on three main steps: (i) optimization of the different factors affecting sample preparation; (ii) evaluation of the main quality parameters of the proposed method; and (iii) application of the method to real samples. The factors influencing the extraction have been optimized by one-at-a-time conventional method (Chapter 3 and Section III.4.3) or by multivariate


338

optimization (other Chapters/Sections). This latter strategy has been the most frequently used since it saves time and resources and is able to show interactions between factors. In all cases, a screening step has been performed because of the large number of experimental factors affecting the extraction. Subsequently, CCD has been performed considering the factors with greater influence. CCD has allowed us to obtain the optimum extraction conditions running a reduced number of experiments.

Finally, regarding separation and detection systems, a first clear differentiation can be established between the first three chapters, in which traditional instrumental techniques have been employed (Table IV.9), and Chapter 4, in which SPELs have been proposed as alternative detection systems (Table IV.10). Chapters 1 and 2 have included the analysis by GC-MS. In both cases, SIM mode was selected with the aim of increasing detector sensitivity and reducing matrix signal. In Chapter 1, the eluent phase was directly injected into the chromatographic system whereas, in Chapter 2, thermal desorption has been used as sample introduction system. With this last system analytes were desorbed and addressed into the column while the entry of the MIL was prevented due to its low volatility. LC technique has been used in Chapter 3 to separate the catecholamines used as model analytes prior to detection by UV or MS/MS. The UV detector was employed during the optimization of extraction conditions, due to its higher simplicity and affordability. However, the complexity of urine samples and the need to reach lower LODs made it necessary UHPLC-MS/MS to evaluate the proposed method. Finally, SPELs presented in Chapter 4 have provided an interesting alternative to the chromatographic techniques employed in previous chapters, which are time-consuming and include bulky and expensive instrumentation. The SPELs were based on working electrodes of different composition depending on the analyte under study. In this regard, commercial SPGEs and SPCEs have been used as received without any further pretreatment or surface modification prior to the determination of TNT and hydrophilic phenols, respectively. With these devices, good analytical parameters (e.g., sensitivity, LOD) were obtained at a very low cost. However, the determination of mercury in water and urine samples needed the modification of SPCEs with gold nanoparticles. The resulting SPCnAuEs combined the advantages of gold


339

as electrode material (i.e., high affinity for mercury, reduction at more positive potentials), with the high surface area of nanoparticles, that significantly enhances the analytical response compared to a mass gold electrode.

Table IV.9. Separation-detection systems and analysis time of the analytical methods included in the first three Chapters of this overview.

Table IV.10. Electrodes, electroanalytical techniques and analysis time of the analytical methods included in Chapter 4.

Chapter Analytes Sample Separation-detection system

Analysis time

1 BTEX Water GC-MS 28 min

2 Chlorobenzenes Water TD-GC-MS 45 min

3

DA, E, NE Urine HPLC-UV 20 min

DA, E Urine UHPLC-MS/MS 5.5 min

Section Analytes Sample SPEL (technique) Analysis time

III. 4.1 TNT Water SPGE (DPV)

Purge for 20 min with N2 and scanning potentials in few

seconds

III. 4.2 Mercury Water SPCnAuE (SWASV)

4 min of deposition and scanning potentials in few

seconds

III. 4.3 Mercury Urine SPCnAuE (SWASV)

4 min of deposition and scanning potentials in few

seconds

III. 4.4 Hydrophilic

phenols Oil SPCE

(DPV) Few seconds

V. Conclusiones generales/

General conclusions

Conclusiones generales

343

V.1. Conclusiones generales Las principales conclusiones que se derivan del trabajo de investigación

que se recoge en esta memoria se describen a continuación:

Capítulo 1: el nuevo sorbente sólido magnético presentado, basado en la zeolita sintética ZSM-5 modificada con nanopartículas de Fe2O3, ha permitido la MSPE de BTEX alcanzando unos LODs que satisfacen la normativa vigente sobre el contenido de estos compuestos en aguas destinadas al consumo. Este nuevo sorbente presenta propiedades destacables como bajo coste y amplia disponibilidad, síntesis sencilla, fácil manipulación y opciones de reutilización. Por tanto, se considera que su utilización para la extracción de diferentes analitos tanto de naturaleza orgánica como inorgánica (i.e., metales y compuestos organometálicos) posee un futuro muy prometedor.

Capítulo 2: se ha presentado un nuevo MIL hidrofílico como fase extractante para la extracción de clorobencenos desde de muestras acuosas mediante la técnica de HS-SDME empleando un pequeño imán de Nd como soporte. La utilización de un sistema de TD ha permitido la determinación de los analitos extraídos en el MIL mediante GC-MS. Los parámeros analíticos obtenidos son comparables a los obtenidos con otros ILs y sistemas análogos. Sin embargo, el sistema propuesto presenta las ventajas del pequeño volumen de MIL utilizado y su fácil manipulación.

Capítulo 3: se ha conseguido por primera vez la EME de especies básicas con una elevada polaridad utilizando un sistema de formación de complejos selectivos y reversibles en la interfase muestra/SLM. Aunque el trabajo presentado tiene un carácter preliminar, los resultados obtenidos son muy prometedores y abren una nueva vía de investigación sobre la EME de especies polares mediada por complejación.

Capítulo 4: se han desarrollado diferentes métodos analíticos en los que se utilizan sistemas miniaturizados tanto en la etapa de preparación de la muestra como en la etapa de detección. La técnica DLLME, en todas sus modalidades, es de manejo sencillo, rápida, económica, ecológica y eficaz,

Conclusiones generales

344

ya que se logra una buena extracción de los analitos en un periodo corto de tiempo por el buen contacto existente entre las fases. Por su parte, los SPELs ofrecen una buena sensibilidad y una rápida respuesta con una instrumentación relativamente económica, sencilla, fácil de manejar y con opciones de portabilidad. Todos los métodos desarrollados poseen unos parámetros analíticos comparables, o incluso mejores, que los proporcionados por otros métodos para los mismos fines.

De forma general, en el presente trabajo se han desarrollado y evaluado nuevas estrategias para la miniaturización de la preparación de la muestra y de sus sistemas de detección electroquímicos. De este modo, se ha mejorado la etapa más crítica del proceso analítico total y se ha avanzado en el desarrollo de métodos analíticos en formato miniaturizado, rápidos y portátiles.

General conclusions

345

V.2. General conclusions

The main general conclusions of the research work presented in this overview are described below:

Chapter 1: a new magnetic solid sorbent based on the synthetic ZSM-5 zeolite modified with Fe2O3 nanoparticles has been introduced for the MSPE of BTEX from water samples. The proposed method reached LOD values that satify the current normative for BTEX content in waters intended for human consumption. The new composite has remarkable properties such as low-cost and availability, simple synthesis, easy manipulation and reuse options. Therefore, it is considered that the application for this new sorbent for the extraction of different analytes, both organic and inorganic (i.e., metals and organometals), has a very promising future.

Chapter 2: a new hydrophilic MIL has been presented for the extraction of chlorobenzenes from water samples using HS-SDME. A small Nd magnet has been used to maintain the MIL drop on the headspace of the extraction vial. The use of a TD system has allowed the determination of analytes extracted with the MIL by GC-MS. The analytical figures of merit obtained with the proposed methodology are comparable to those obtained with other ILs and analogous systems. However, the suggested experimental setup shows the advantages of using small volumes of MIL and the easy manipulation under a magnetic field.

Chapter 3: EME of highly polar basic analytes has been demonstrated for the first using a selective and reversible complexation process in the sample/SLM interface. Although this work is preliminary in nature, complexation mediated EME shows potential and may open new and very interesting future possibilities for EME of polar species.

Chapter 4: different analytical methods have been developed using miniaturized systems both in sample preparation and detection stage. On the one hand, the different modalities of DLLME are simple, fast,

General conclusions

346

economical, ecological and efficient, since a great extraction of the analytes is achieved with short extraction times due to the good contact between phases. On the other hand, SPELs offer good sensitivity and rapid response with simple, inexpensive and portable instrumentation. In all the developed methods, the analytical figures of merit are comparable or even better than those obtained by other methods suggested for the same purpose.

In general, new strategies have been developed for the miniaturization of sample preparation and its electrochemical detection systems. Thus, the most critical stage of the total analytical process has been improved and the development of miniaturized, fast and de-centralized analytical methods has been progressed.

VI. Trabajos futuros/Future works

Trabajos futuros

351

VI.1. Trabajos futuros Con el trabajo experimental presentado en esta memoria se han realizado

aportaciones significativas en el campo de preparación de la muestra con fines analíticos. Sin embargo, durante el desarrollo del mismo también se han planteado nuevos interrogantes o retos que es necesario abordar. Entre los más significativos se encuentran:

Estudiar diferentes tipos de zeolitas, tanto de origen natural como sintético, así como su modificación superficial para mejorar el rendimiento de las extracciones.

Utilizar MOFs magnéticos como fase extractante en la técnica de MSPE y comparar su capacidad de extracción con la de las zeolitas magnéticas.

Investigar nuevas SLMs para conseguir la EME eficiente y estable de especies polares desde muestras biológicas como la orina, el plasma o la sangre.

Sintetizar, caracterizar y aplicar nuevos MILs al tratamiento de muestas acuosas con el fin de superar los inconvenientes que todavía presentan estos disolventes con propiedades únicas.

Combinar diferentes técnicas de SPME (i.e., DSPME, MSPE) basadas en zeolitas con la detección electroquímica con SPELs, de manera que tras la extracción, el sólido sorbente enriquecido se deposite directamente sobre la superficie del electrodo y sea posible la detección de los analitos sin llevar a cabo ninguna etapa de elución.

Aplicar el método desarrollado en la Sección III.4.4 a un mayor número de muestras con el fin de establecer si la metodología propuesta es capaz de detectar adulteraciones o fraudes en el etiquetado de aceites de oliva comerciales.

Future works

352

VI.2. Future works The experimental work presented in this overview has provided significant

contributions in the sample preparation field, but new questions or challenges have also come up that need to be addressed. Among them, we can mention:

Study different types of zeolites, both natural and synthetic, and their surface modification in order to improve extraction yields.

Use magnetic MOFs as extractant phase in MPSE and compare the extraction capacity with magnetic zeolites.

Investigate new SLMs for the efficient and stable EME of polar species from biological samples such as urine, plasma or blood.

Synthesize, characterize and apply new MILs to the treatment of aqueous samples in order to overcome the disadvantages that these solvents with unique properties still possess.

Combine different SPME techniques (i.e., DSPME, MSPE) based on zeolites with the electrochemical detection with SPEL so that, after extraction, the enriched solid sorbent is directly deposited on the electrode surface allowing the determination of analytes without any further dilution steps.

Apply the method developed in Section III.4.4 to a greater number of samples in order to establish if it is able to detect adulterarions or frauds in the labeling of commercial olive oils.

VII. ANEXOS

Anexos

357

VII.1. Comunicaciones en congresos

19th International Symposium on Advances in Extraction technologies, Santiago de Compostela (España), junio 2017

Comunicación oral: Complexation mediated electromembrane extraction of highly polar basic drugs. E Fernández, L. Vårdal, L. Vidal, A. Canals, A. Gjelstad, S. Pedersen-Bjergaard. Comunicación oral: Magnetic headspace adsorptive extraction of chlorobenzenes prior to thermal desorption gas chromatography-mass spectrometry. L. Vidal, M. Ahmadi, E. Fernández, T. Madrakian, A. Canals.

XXXVII Reunión del Grupo de Electroquímica, Alicante (España), julio 2016

Póster: Determinación de fenoles hidrofílicos en aceites de oliva mediante microextracción líquido-líquido dispersiva y electrodos serigrafiados: una alternativa al método Folin-Ciocalteu. E. Fernández, L. Vidal, A. Canals.

XXV National Spectroscopy Meeting and IX Iberian Spectroscopy Conference, Alicante (España), julio 2016

Póster: Magnetic solid-phase extraction using ZSM-5 zeolite for BTEX determination in water samples. E. Fernández, L. Vidal, A. Canals.

XVII Euroanalysis. European Conference on Analytical Chemistry, Burdeos (Francia), septiembre 2015

Póster: Mercury determination in urine samples by dispersive liquid-liquid microextraction and screen-printed electrodes. E. Fernández, L. Vidal, A. Costa, A. Canals.

Comunicaciones en congresos

358

14as Jornadas de Análisis Instrumental, Barcelona (España), octubre 2014

Póster: Mercury determination in biological samples by dispersive liquid-liquid microextraction and screen-printed electrodes. E. Fernández, L. Vidal, A. Canals, A. Costa.

XXXV Meeting of Electrochemistry of the Spanish Royal Society of Chemistry, Burgos (España), julio 2014

Póster: Electrochemical determination of mercury in environmental and drinking waters by using ionic liquid microextraction. E. Fernández, L. Vidal, D. Martín, M.C. Blanco,

A. Canals, A. Costa.

16th International Symposium on Advances in Extraction technologies, Chania (Grecia), mayo 2014

Comunicación oral: Screen-printed electrode-based electrochemical detector coupled with dispersive liquid-liquid microextraction for determination of mercury in water samples. E. Fernández, L. Vidal, D. Martín, M.C. Blanco, A. Canals, A. Costa. Comunicación oral: Non conventional detectors for liquid-phase microextraction. L. Vidal, E. Fernández, A. Canals.

XVIII Reunión Nacional de la Sociedad Española de Química Analítica, Úbeda (España), julio 2013

Póster: In-situ ionic liquid dispersive liquid-liquid microextraction of 2,4,6-trinitrotolueno in water samples. E.J. Ródenas, E. Fernández, A. Hristeva, L. Vidal, A. Canals.

Anexos

359

5th International Congress on Ionic Liquids (COIL-5), Algarve (Portugal), abril 2013

Comunicación oral: Can ionic liquid dispersive liquid-liquid microextraction be combined with screen-printed electrodes as electrochemical detectors? E. Fernández, L. Vidal, J. Iniesta, C.E. Banks, A. Canals.

Otras publicaciones

360

VII.2. Otras publicaciones

Capítulos de libros

E. Fernández, L. Vidal, Liquid phase extraction and microextraction, en: Ionic Liquids in Separation Technology, Elsevier Science, 2014: pp. 107-152. ISBN: 978-0-444-63257-9.

E. Fernández, L. Vidal, Liquid-phase microextraction techniques, en: Miniaturization in Sample Preparation, De Gruyter Open, 2014: pp. 191-252. ISBN 978-3-11-0410181.

Artículos en revistas

L. Vidal, M. Ahmadi, E. Fernández, T. Madrakian, A. Canals, Magnetic headspace adsorptive extraction of chlorobenzenes prior to thermal desorption gas chromatography-mass spectrometry, Analytica Chimica Acta 971 (2017) 40-47.

A. Diuzheva, E. Fernández, L. Vidal, V. Andruch, A. Canals, Vitamin E determination in edible oils by reversed-phase dispersive liquid-liquid microextraction and screen-printed carbon electrodes, (enviado a RSC Advances).

nuevos avances en metodologías analíticas basadas en...

Documents